EVALUATING THE POTENTIAL FOR
INCREASED FORAGE PRODUCTIVITY AND
SOIL CARBON SEQUESTRATION IN STRIPTHINNED SILVOPASTURES.
By
STEVEN KEGA
Thompson Rivers University, 2021

A THESIS SUBMITTED IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR
THE DEGREE OF

MASTER OF SCIENCE in ENVIRONMENTAL SCIENCE
in the Faculty of Science

Thesis examining committee:
Dr. Lauchlan Fraser (PhD), Professor and Thesis Supervisor, Department of Natural Resource
Sciences, Thompson Rivers University.

Dr. John Church (PhD), Associate Professor and Thesis Co-Supervisor, Department of Natural
Resource Sciences, Thompson Rivers University.

Dr. John Karakatsoulis (PhD), Associate Teaching Professor and Committee Member,
Department of Natural Resource Sciences, Thompson Rivers University.

Steven Kega, 2021

i

ABSTRACT
Forested ecosystems are essential in supporting the majority of terrestrial species on Earth.
Forest products contribute significantly to the global economy as well as contributing in
sequestering carbon as part of climate change mitigation. In British Columbia, conventional forest
and range management has historically considered multiple-use landscape resources
independently. We explored an opportunity to integrate the forest and ranching industries, in order
to enhance both forestry and grazing practices, so that forest production and understory forage
productivity can be fully realized. Silvopasture, which is the complementary use of land for
forestry and range productivity for livestock, is a practice that integrates these two sectors.
Previous research has shown that a successful integration of forage, cattle and timber management
can provide significant economic, social and environmental benefits such as increasing forage
yield and quality, tree growth, enhancing soil carbon storage, and increasing soil water availability.
Our objective was to test the integration of forage and timber management to improve forage
quantity, quality and to enhance soil carbon sequestration. Specifically, we tested two hypotheses:
(H1) 20 m width strip–thinning will maximize forage yield and quality and (H2) 20 m width
thinning will sequester more soil carbon than uncut control or 10 m and 15 m thinning. In British
Columbia, Canada, a mid-rotation forest of planted 45-year old lodgepole pine was harvested July
2018 at 10 m, 15 m and 20 m width strips in three adjacent forest sites at an elevation range
between 1340 – 1400 m. Baseline data including tree stand density, understory plant species
composition, and soil carbon and nitrogen were collected pre-harvest, June 2018. An agronomic
seed mix was broadcast at 12 kg/ha: 30% Dactylis glomerata, 30% Bromus riparius, 30%
Thinopyrum intermedium, and 10% Trifolium repens was seeded in October 2018. Field
experiments, laboratory analysis and remote sensing were used in the second and third phase of
this research to monitor forage quality and quantity, soil total carbon, nitrogen and organic carbon
as well as soil carbon sequestration. We found that all strip widths enhanced forage quality, but
the 20 m strips produced more yield than other treatment units. We found higher soil compaction
and increased pH level in 20 m strips than other treatment units. However, a higher soil carbon
and nitrogen was found in 15 m and 10 m strips than in 20 m. Our results provide evidence for
optimizing land use in silvopasture.

ii
Keywords: forested ecosystem, forest and range productivity, climate change, carbon
sequestration, silvopasture, multiple use, ecosystem goods and services.

iii
TABLE OF CONTENTS

ABSTRACT ................................................................................................................................ i
TABLE OF CONTENTS ......................................................................................................... iii
ACKNOWLEDGEMENTS ..................................................................................................... vi
LIST OF FIGURES ................................................................................................................ vii
LIST OF TABLES ................................................................................................................. xiii
LIST OF ACRONYMS .......................................................................................................... xiv
CHAPTER 1- GENERAL INTRODUCTION ........................................................................... 1
Grazing Management ............................................................................................................ 2
Ecological restoration and carbon sequestration through silvopastoralism .................... 3
Remote sensing technology in silvopasture systems ........................................................... 5
Research objectives................................................................................................................ 7
LITERATURE CITED ............................................................................................................. 9
CHAPTER 2 – UNDERSTORY PLANT COMMUNITY ESTABLISHMENT WITHIN A
LODGEPOLE PINE SILVOPASTURE SYSTEM ................................................................. 15
INTRODUCTION ............................................................................................................... 15
MATERIALS & METHODS ............................................................................................. 18
Site Description .................................................................................................................... 18
Research Design ................................................................................................................... 19
Pre-harvest field data collection ......................................................................................... 20
Post-harvest field data collection........................................................................................ 21
Remote sensing data collection ........................................................................................... 22
Remote sensing data processing ......................................................................................... 24
STATISTICAL ANALYSIS ............................................................................................... 24

iv
RESULTS ............................................................................................................................. 25
Forage productivity ............................................................................................................. 25
Yield ...................................................................................................................................... 25
Quality .................................................................................................................................. 26
Species richness, and diversity ........................................................................................... 28
Understory community composition .................................................................................. 31
Strip thinned NDVI and NDRE index products derived from multispectral remote
sensing ................................................................................................................................... 41
DISCUSSION ....................................................................................................................... 43
Forage yield and quality...................................................................................................... 43
Understory species richness and diversity......................................................................... 44
Understory species community composition ..................................................................... 44
NDVI and NDRE vegetation index products .................................................................... 45
CONCLUSION .................................................................................................................... 45
LITERATURE CITED ........................................................................................................... 47
CHAPTER 3 – EFFECT OF SILVOPASTURE SYSTEMS ESTABLISHED THROUGH
FOREST THINNING ON SOIL CARBON AND NITROGEN STOCKS. .......................... 54
INTRODUCTION ............................................................................................................... 54
MATERIALS & METHODS ............................................................................................. 57
Site Description & Research Design .................................................................................. 57
Soil sampling for bulk density analysis ............................................................................. 57
Soil sampling for pH, SOM, TN & TC analysis ................................................................ 58
Soil pH .................................................................................................................................. 58
Soil Organic Matter (SOM) ................................................................................................ 59
Soil total nitrogen (TN) and total carbon (TC) ................................................................. 59
Net Carbon Exchange (NCE) through soil respiration .................................................... 60

v
STATISTICAL ANALYSIS ............................................................................................... 61
RESULTS ............................................................................................................................. 62
Soil bulk density ................................................................................................................... 62
Soil pH .................................................................................................................................. 63
SOM ...................................................................................................................................... 65
Total Carbon (TC) ............................................................................................................... 68
Total Nitrogen (TN) ............................................................................................................. 70
Net Carbon Exchange (NCE) ............................................................................................. 72
DISCUSSION ....................................................................................................................... 75
Effect of strip thinning on soil physical and chemical properties ................................... 75
Bulk density and pH ............................................................................................................ 75
Soil Organic Matter (SOM) ................................................................................................ 76
Total Carbon (TC) and Nitrogen (TN) .............................................................................. 76
Net Carbon Exchange (NCE) ............................................................................................. 77
CONCLUSION .................................................................................................................... 78
LITERATURE CITED ........................................................................................................... 79
CHAPTER 4 – MANAGEMENT IMPLICATIONS AND FUTURE RESEARCH ............ 86
LITERATURE CITED ........................................................................................................... 87
APPENDIX – A ....................................................................................................................... 88
Relationships between understory species community in the thinned areas Pre-harvest
(July 2018) and Post-harvest (August 2019 and July 2020). ............................................ 88

vi
ACKNOWLEDGEMENTS
I would like to sincerely thank my thesis committee:
1. My thesis supervisor, Dr. Lauchlan Fraser for his exceptional guidance and support
throughout the course of my project. Thank you for your unbelievable encouragement and
patience from the beginning of my program when I had no idea of what to do or how to do
it.
2. Dr. John Church for his expertise, patience and assistance throughout my field data
collection. Thank you for always being available for help, for all the great moments in the
field during nice sunny and rainy days, and for always reminding me to SECURE THE
DATA!
3. Dr. John Karakatsoulis for his excellent suggestions and advice throughout this MSc
program. Thank you for your help and guidance during my upgrade year in natural resource
program.
4. Dr. Lisa Zabek for being available anytime time I needed guidance in the field. Thank you
for constant advice and support with data collection.
I would like to thank everyone at the Fraser Lab and research assistants that have contributed to a
friendly and educational working and learning environment. I am so thankful for your
encouragement and assistance with field work. I would also like to thank my parents and Canadian
families – Heney family, Edgeworth family - for their loving support, I would not be where I am
today without their help.
I would also like to acknowledge the support provided by BCRC (Beef Cattle Research Council),
BCCA (British Columbia Cattlemen’s Association), TOLKO Industries Ltd, BC Ministry of
Agriculture, FLNRORD (Ministry of Forests, Lands, Natural Resource Operations and Rural
Development) and NSERC (Natural Sciences and Engineering Research Council).

vii
LIST OF FIGURES
Figure 1.1 “Soil carbon in and output by plants and associated soil heterotrophs. Solid lines
indicate carbon incorporation and dotted lines are soil carbon loss; SOC: soil organic carbon, VOC:
volatile organic carbon, DOC: dissolved organic carbon” (De Deyn et al. 2008).......................... 5
Figure 1.2: An example of a typical plant reflectance curve detailing the wavelengths and position
of visible and non-visible light (Micasense Inc.) (Parker 2019). .................................................... 7
Figure 2.1: An operational- scale pilot map initiated at a 101.4 ha area in the Okanagan region,
Goudie, Kelowna, BC. The experiment was conducted on three blocks with four treatments. The
red marked areas are control treatments. The green strips are 10 m wide, purple strips are 15 m
wide and blue strips 20 m width. The yellow pins indicate the sampling points. ........................ 19
Figure 2.2. An extended Daubenmire method of a 0.25 sq. m quadrat used to collect % cover data
per species within three adjacent forest sites located in Goudie, Kelowna, British Columbia before
strip harvesting activities. ............................................................................................................. 21
Figure 2.3. Drone (UAV currently known as RPA) equipped with a high resolution multispectral
camera with a blue (475 nm center, 20 nm bandwidth), green (560 nm center, 20 nm bandwidth),
red (668 nm center, 10 nm bandwidth), red edge (717 nm center, 10 nm bandwidth), near-IR (840
nm center, 40 nm bandwidth) narrow bands and thermal sensor (LWIR: 8- 14μm). ................... 23
Figure 2.4. D-RTK 2 mobile station (left), a high-precision global navigation satellite system
receiver compatible with a new aircraft version, it uses navigational satellites from other networks
in providing Aircraft accuracy. Multispectral sensor calibrated reflectance panel (right). .......... 23
Figure 2.5 Mean total biomass weight in gram per 0.25 sq.m for understory vegetation harvested
in each treatment in August 2019 and July 2020. Error bars represent standard error of the mean
(n= 6 for each group in 2019 and 30 for each group in 2020). (NS= non-significant; * = p≤ 0.05;
** = p≤ 0.01; *** = p≤ 0.001; **** = p≤ 0.0001). ....................................................................... 26
Figure 2.6 Mean species richness for understory vegetation harvested in each treatment unit preharvest 2018, post-harvest 2019 and 2020. Error bars represent standard error of the mean (n= 30

viii
for each group). (NS= non-significant; * = p≤ 0.05; ** = p≤ 0.01; *** = p≤ 0.001; **** = p≤
0.0001). ......................................................................................................................................... 29
Figure 2.7 Shannon-Wiener (H) mean species diversity for understory vegetation harvested in
each treatment unit pre-harvest 2018, post-harvest 2019 and post-harvest 2020. Error bars
represent standard error of the mean (n= 30 for each group). (NS= non-significant; * = p≤ 0.05; **
= p≤ 0.01; *** = p≤ 0.001; **** = p≤ 0.0001). ............................................................................ 30
Figure 2.8 PCoA (Principal Coordinates Analysis) ordination method for year 2018 pre-harvest
species community composition of four treatment units in the three adjacent forest areas. ........ 33
Figure 2.9 PCoA for year 2018 pre-harvest species community composition comparing uncut
control and 10 m width treatments in the three adjacent forest blocks. ........................................ 33
Figure 2.10 PCoA comparison between uncut control and 15 m width treatments in the three
adjacent forest blocks pre-harvest June 2018. .............................................................................. 34
Figure 2.11 PCoA comparison between uncut control and 20 m width treatments in the three
adjacent forest blocks pre-harvest June 2018. .............................................................................. 34
Figure 2.12 PCoA comparing understory species community composition between Uncut control
treatment, 10 m, 15 m and 20 m strip thinned treatments one year post timber harvest (August
2019). ............................................................................................................................................ 36
Figure 2.13 PCoA analysis comparing uncut control treatment to 10 m, 15 m and 20 m thinned
treatments individually after one-year post harvest (August 2019). ............................................. 37
Figure 2.14 PCoA comparing understory species community composition across all treatment
units two year post timber harvest (July 2020). ............................................................................ 39
Figure 2.15 PCoA comparing understory species community composition between uncut control
and 10 m thinned treatments two year post timber harvest (July 2020). ...................................... 39
Figure 2.16 PCoA comparing understory species community composition between uncut control
and 15 m thinned treatments two year post timber harvest (July 2020). ...................................... 40

ix
Figure 2.17 PCoA comparing understory species community composition between uncut control
and 20 m thinned treatments two year post timber harvest (July 2020). ...................................... 40
Figure 2.18 Orthomosaic images with corresponding NDVI and NDRE indexes of the three
adjacent forest sites located in Goudie, East Kelowna, British Columbia. Orthoimages and
multispectral sensor images were taken post timber harvest. ....................................................... 41
Figure 3.1: “Soil organic carbon (SOC); a) and soil organic nitrogen (SON); b) controls at
different temporal scales. Dashed lines show which controls are affected by grazing. ANPP is
aboveground net primary production, and BNPP is belowground net primary production” (Pineiro
et al. 2010). ................................................................................................................................... 56
Figure 3.2: Thermo Fisher FlashSmart Elemental Analyzer Nitrogen and Carbon configuration
and analysis process (Krotz et al. 2016). ...................................................................................... 60
Figure 3.3: LI-COR, LI-8100A automated CO2 soil gas flux system with a direct survey chamber
ranged between 0ppm to 20,000ppm and an operating temperature range between -20oC to 45oC.
A green soil Collar with a 10 cm inside and 11.4 cm outside diameter, and with 10 cm of length
inserted permanently in the soil, and a minimum of 2 cm extension above the soil surface (Li-COR
Biosciences 2010a). ...................................................................................................................... 61
Figure 3.4: Post-harvest 2019 and 2020 mean bulk density (BD in g cm-3) of each treatment unit
in three adjacent block design. Error bars represent standard error of the mean (n= 6 for each group
in 2019 and 48 in May 2020). (NS= non-significant; * = p≤ 0.05; ** = p≤ 0.01; *** = p≤ 0.001;
**** = p≤ 0.0001). ........................................................................................................................ 63
Figure 3.5: Mean soil pH above 10 cm depth of each treatment unit in three adjacent block designs
pre-harvest 2018, post-harvest 2019 and 2020. Error bars represent standard error of the mean
(n=30 for each group). (NS= non-significant; * = p≤ 0.05; ** = p≤ 0.01; *** = p≤ 0.001; **** =
p≤ 0.0001). .................................................................................................................................... 64
Figure 3.6: Mean soil pH below 10 cm depth of each treatment unit in three adjacent block designs
pre-harvest 2018, post-harvest 2019 and 2020. Error bars represent standard error of the mean

x
(n=30 for each group). (NS= non-significant; * = p≤ 0.05; ** = p≤ 0.01; *** = p≤ 0.001; **** =
p≤ 0.0001). .................................................................................................................................... 65
Figure 3.7 Mean SOM (Soil Organic Matter) above 10 cm depth of each treatment unit in three
adjacent block designs pre-harvest 2018, post-harvest 2019 and 2020. Error bars represent standard
error of the mean (n=30 for each group). (NS= non-significant; * = p≤ 0.05; ** = p≤ 0.01; *** =
p≤ 0.001; **** = p≤ 0.0001). ....................................................................................................... 67
Figure 3.8: Mean SOM (Soil Organic Matter) below 10 cm up to 20 cm soil depth of each
treatment unit in three adjacent block designs pre-harvest 2018, post-harvest 2019 and 2020. Error
bars represent standard error of the mean (n=30 for each group). (NS= non-significant; * = p≤
0.05; ** = p≤ 0.01; *** = p≤ 0.001; **** = p≤ 0.0001)............................................................... 68
Figure 3.9: Mean TC (Total Carbon) above 10 cm soil depth of each treatment unit in three
adjacent block designs pre-harvest 2018, post-harvest 2019 and 2020. Error bars represent standard
error of the mean (n=30 for each group). (NS= non-significant; * = p≤ 0.05; ** = p≤ 0.01; *** =
p≤ 0.001; **** = p≤ 0.0001). ....................................................................................................... 69
Figure 3.10: Mean TC (Total Carbon) below 10 cm up to 20 cm soil depth of each treatment unit
in three adjacent block designs pre-harvest 2018, post-harvest 2019 and 2020. Error bars represent
standard error of the mean (n=30 for each group). (NS= non-significant; * = p≤ 0.05; ** = p≤ 0.01;
*** = p≤ 0.001; **** = p≤ 0.0001). ............................................................................................. 70
Figure 3.11: Mean TN (Total Nitrogen) above 10 cm soil depth of each treatment unit in three
adjacent block designs pre-harvest 2018, post-harvest 2019 and 2020. Error bars represent standard
error of the mean (n=30 for each group). (NS= non-significant; * = p≤ 0.05; ** = p≤ 0.01; *** =
p≤ 0.001; **** = p≤ 0.0001). ....................................................................................................... 71
Figure 3.12: Mean TN (Total Nitrogen) below 10 cm up to 20 cm soil depth of each treatment
unit in three adjacent block designs pre-harvest 2018, post-harvest 2019 and 2020. Error bars
represent standard error of the mean (n=30 for each group). (NS= non-significant; * = p≤ 0.05; **
= p≤ 0.01; *** = p≤ 0.001; **** = p≤ 0.0001). ............................................................................ 72

xi
Figure 3.13: Mean CO2 Flux content of the treatment units in three adjacent study blocks collected
in May 2019, July 2019 and October 2019. Error bars represent standard error of the mean (n= 6
for each group). (NS= non-significant; * = p≤ 0.05; ** = p≤ 0.01; *** = p≤ 0.001; **** = p≤
0.0001). ......................................................................................................................................... 73
Figure 3.14: Mean CO2 Flux content of the treatment units in three adjacent study blocks collected
in May 2020, July 2020 and October 2020. Error bars represent standard error of the mean (n= 6
for each group). (NS= non-significant; * = p≤ 0.05; ** = p≤ 0.01; *** = p≤ 0.001; **** = p≤
0.0001). ......................................................................................................................................... 74
Figure A.1. PCoA analysis based on Bray-Curtis dissimilarity for July 2018 understory species
community composition between areas assigned for 10 m and 15 m treatment widths. .............. 88
Figure A.2. PCoA analysis based on Bray-Curtis dissimilarity for July 2018 understory species
community composition between areas assigned for 10 m and 20 m treatment widths. .............. 88
Figure A.3. PCoA analysis based on Bray-Curtis dissimilarity for July 2018 understory species
community composition between areas assigned for 15 m and 20 m treatment widths. .............. 89
Figure A.4. PCoA analysis based on Jaccard dissimilarity for August 2019 understory species
community composition between 10 m and 15 m thinned treatments.......................................... 89
Figure A.5. PCoA analysis based on Jaccard dissimilarity for August 2019 understory species
community composition between 10 m and 20 m thinned treatments.......................................... 90
Figure A.6. PCoA analysis based on Jaccard dissimilarity for August 2019 understory species
community composition between 15 m and 20 m thinned treatments.......................................... 90
Figure A.7. PCoA analysis based on Jaccard dissimilarity for July 2020 understory species
community composition between 10 m and 15 m thinned treatments.......................................... 91
Figure A.8. PCoA analysis based on Jaccard dissimilarity for July 2020 understory species
community composition between 10 m and 20 m thinned treatments.......................................... 91

xii
Figure A.9. PCoA analysis based on Jaccard dissimilarity for July 2020 understory species
community composition between 15 m and 20 m thinned treatments.......................................... 92
A.10. Soil Horizons, texture and coarse fragment determination pre-timber harvest in three
adjacent block design. ................................................................................................................... 92
A.11. Soil texture diagram of Goudie silvopasture research area contains 46.4% sand, 47.9% silt
and 5.8% clay. ............................................................................................................................... 93
A.12. Excavation method and materials for soil bulk density analysis used post-harvest 2020. . 94
A.13. An NDVI map example that shows soil compaction in the strips post timber harvesting. 94
A.14. Integrated forage, livestock and timber approach through forest strip thinning. ................ 95
A.15: Map of one of the experimental design blocks detailing the strip widths and their timber
buffers. .......................................................................................................................................... 96

xiii
LIST OF TABLES
Table 2.1. Highly palatable agronomic plant species used for improving the quantity and quality
of animal forage ............................................................................................................................ 20
Table 2.2: Post timber harvest 2019 and 2020 mean understory seeded agronomic vegetation
harvested in the opening strips. ..................................................................................................... 27
Table 2.3: Post timber harvest 2019 and 2020 mean understory native vegetation harvested in all
treatment units. .............................................................................................................................. 28
Table 2.4: ANCOVA and post-hoc test results for understory species richness and diversity preharvest 2018 and post-harvest 2019-2020. ................................................................................... 31
Table 2.5. Estimate percent cover of six native grazed plant species on three experimental design
blocks pre-harvest 2018 in Goudie, Kelowna B.C ....................................................................... 32
Table 2.6. Estimate percent cover of thirteen grazed plant species on three experimental design
blocks post-harvest 2019 in Goudie, Kelowna B.C ...................................................................... 35
Table 2.7. Estimate percent cover of fifteen grazed plant species on three experimental design
blocks post-harvest 2020 in Goudie, Kelowna B.C ...................................................................... 38
Table 2.8: Mean values comparison of understory vegetation greenness and availability measured
by NDVI and NDRE index products derived from a multispectral sensor between strip-thinned
treatments ...................................................................................................................................... 42
Table 3.1: ANCOVA and post-hoc test results for soil total carbon (0-10 cm soil depth) pre-harvest
2018 and post-harvest 2019-2020. ................................................................................................ 72
A.16: Plant species identified from all three blocks in Goudie, Kelowna Pre and post strip thinning.
(USDA- United States Department of Agriculture 2019)............................................................. 97

xiv
LIST OF ACRONYMS
ANPP: Aboveground Net Primary Production
AGL: Above Ground level
ADF: Acid Detergent Fiber
BC: British Columbia
BNPP: Belowground Net Primary Production
CP: Crude Protein
DOC: Dissolved Organic Carbon
DLS: Downwelling Light Sensor
DSM: Digital Surface Model
GDP: Gross Domestic Product
Gt: gigatonne
GPS: Global Positioning System
GSD: Ground Sample Distance
GNSS: Global Navigation Satellite System
LWIR: Long-Wave Infrared
Mt: megatonnes
NCE: Net Carbon Exchange
NIR: Near- Infrared
NDVI: Normalized Difference Vegetation Index
NDRE: Normalized Difference RedEgde
NDF: Neutral detergent Fiber
PCoA: Principal Coordinates Analysis
PERMANOVA: Permutational multivariate analysis of variance
%TN: Percentage Total Carbon
%TN: Percentage Total Nitrogen
RPAS: Remotely Piloted Aircraft system
RTK: Real-Time Kinematic
SON: Soil Organic Nitrogen
SOM: Soil Organic Matter
SOC: Soil Organic Carbon
TDN: Total Digestible Nutrients
UAVs: Unmanned Aerial Vehicles
USDA: United States Department of Agriculture
UTM: Universal Transverse Mercator
VOC: Volatile Organic Carbon
WGS: World Geodetic System

1

CHAPTER 1- GENERAL INTRODUCTION
A significant rise in the human population and high demand in livestock products have led to a
rapid increase in the livestock sector (Onteru et al. 2010). Approximately 30 percent of the earth
surface is occupied by livestock production which generates more than $1.4 trillion of the global
asset (Thornton 2010). The livestock sector has contributed significantly to humanities economic
development in employing over 1.3 billion people globally and impacting the GDP (Gross
Domestic Product) by 33 percent in developing countries (Thornton 2010).
In Canada, this sector has been an important socio-economic driver as it employs around
228,811 people across the country and contributes approximately $13.2 billion to the national GDP
(Kulshreshtha et al. 2012). As a resource-driven economy, it is important to keep developing more
effective techniques to improve and maintain livestock productivity in ways that the sector will
continue to be sustainable socially and economically and keep fostering the integrity of the
ecosystem.
Forest and ranching industries in North America are long-standing practices. However, research
and application of both practices on the same land base is not well understood (Fike et al. 2004).
An intentional management system such as agroforestry has a long history and is recognized as a
sustainable and efficient land-use strategy across the world, in integrating both ranching and forest
industries. The combination of ranching or agriculture with forestry can be beneficial to farmers
in a short and long period of time with the intention of providing significant economic benefits
such as timber, crops, fruits, forage, livestock, and biomass while minimizing nutrient runoff and
soil erosion (Wilson and Lovell 2016).
Over the years, various forms of agroforestry management systems including silvopasture,
alley-cropping, shelterbelts, and buffer strips have been practiced (Wilson and Lovell 2016); but
the silvopasture system has been more often used in many regions, specifically intended and
intensively managed for the production of trees, forage, and livestock (Klopfenstein et al. 1997).
The silvopasture system, an intentional combination of trees along with vegetation and livestock
on the same land base, has historically been recognized as a successful practice with an overall
goal of increasing the total production and benefits of farmers (Verma et al., 2017). Implementing
silvopasture systems may be more beneficial than a single conventional forestry system and a
monoculture forage based grazing system (Frey et al. 2012). Despite the management complexity,

2
silvopasture systems can offer various social, economic and environmental benefits (Jose 2009)
including soil erosion prevention, soil fertility improvement by fixing nitrogen through trees and
forage species, water quality improvement, improved wildlife habitat and biodiversity
conservation (Shrestha et al. 2004; Baah-Acheamfour et al. 2015). In addition, silvopasture
management contributes to carbon sequestration and return on investment by creating a stable
source of cash flow before and after timber harvest (Klopfenstein et al. 1997; Husak and Grado
2002).
The adoption of silvopasture has been highly recommended by several researchers and
extension agents as it can be used on both small and large scales (Frey et al. 2012) in meeting high
global demand in forage productivity for not only livestock, but also food for human consumption
as well as wildlife, with a significant contribution toward environmental accountability. However,
all these benefits will depend on different factors including the management approaches of both
integrated systems (Jose and Dollinger 2019).
Grazing Management
Grazing management is one of the factors that can be used to influence productivity and
sustainability of a silvopasture system. Effective livestock management and understanding animalpasture relationships (stocking rate, stock density, timing and uniformity of use) (Krzic et al. 2004)
by adopting various strategies such as rotational grazing, can be an essential tool that can help in
improving forage productivity, facilitating better plant regrowth after a grazing period as well as
improve soil organic carbon (Sanderman et al. 2015). Rotational grazing is an organized strategy
of moving livestock through different areas to allow forage to restore its reserved energy and
encourage better recovery without diminishing animal performance (Hao et al. 2013; Wang et al.
2020). Rotational grazing is a very important strategy as forage productivity tends to decline with
intensive uncontrolled livestock grazing or severity of weed spread (De Bruijn and Bork 2006;
Hao et al. 2013). In addition, livestock behaviour, timing and grazing duration are important
strategies to consider in enhancing forage productivity and ensuring a sustainable silvopasture
system (Delcurto et al. 2005; Zampaligré and Schlecht 2018) as livestock tend to select more
palatable young forage species than other species; which tends to increase the vigour of unpalatable
species and decrease the growth of palatable forage (Willms et al. 1980).
Forage production in a silvopasture system is not only impacted by improper grazing
management but also the amount of tree canopy closure, variation in soil moisture and solar

3
radiation intercepted and used by the understory forage species in complex biochemical pathways
(Lin et al. 1999). The effect of solar radiation on forage productivity might depend on the types of
understory vegetation species such as cool-season grasses, which tend to be more shade tolerant,
and warm-season grasses, which are considered as low shade tolerant due to different
photosynthetic rates (Lin et al. 1999). In addition, the variation in soil moisture competition
between trees and underlying vegetation (shrubs, forbs and grass) might depend on different soil
depth as moisture variation is often seen above 30 cm soil depth (Jose et al. 2000).
According to Lindgren & Sullivan (2014), thinning is one of the silviculture methods used to
improve forage productivity by increasing soil moisture, and reduce soil nutrients competition,
improving solar radiation and humidity while achieving the ecological goal of promoting
biodiversity. This integration of forage, timber, and livestock on the same land base through forest
thinning can also be economically attractive to farmers and landowners, and is successfully
practiced in different regions around the world (Karki et al. 2009). However, there is still a gap in
scientific studies on how different methods of forest thinning may affect forage quality and
quantity, as well as the quality of the soil. A variation of commercial thinning is strip thinning,
where harvest is concentrated in strips of variable widths. An assessment of strip thinning in
enhancing forage and timber productivity in mid-rotation lodgepole pine using various strip widths
is the primary focus of this study. In addition, this research project also explored the effect of
introducing silvopasture through strip thinning on soil carbon and nitrogen storage.
Ecological restoration and carbon sequestration through silvopastoralism
Ecosystems have been altered due to various causes including human activities with a
significant impact on animal habitats, plant diversity and the introduction of invasive species
(Norris 2012). Ecological restoration is the process of conserving or supporting ecosystem
biodiversity that has been disturbed or damaged, such that a self-sustaining ecosystem is returned
to its natural state in providing ecosystem services (Harris et al. 2006). This concept is considered
one of the effective solutions for habitat recovery, mitigating threatened or endangered species, as
well as a tool to mitigate the effect of climate change (Harris et al. 2006).
The adoption of reforestation or the use of forest management systems in supporting ecological
restoration has contributed positively to the global economy, sequestration of carbon as part of
climate change mitigation and the provision of vital services to the majority of terrestrial species

4
(Seely et al. 2002; Hogarth et al. 2013); including food and shelter for various animals, fuel and
nutrient cycling, timber production and recreation activities (Hall et al. 2019).
Between 1990- 2007, The world’s forests sequestered more than 30% of the total annual green
house gas emissions including atmospheric carbon (Hoberg et al. 2016). An estimated 500 Gt to
800 Gt of carbon sequestered by forest ecosystem was aboveground (Nilsson and Schopfhauser
1995), while most carbon sequestered by grassland ecosystems is found belowground (Soussana
et al. 2004). It is well known that the largest terrestrial carbon sink occurs in a forested ecosystem,
but managed forests through a proper planned agroforestry system has the potential to sequester
more carbon both above and belowground and improve carbon dynamics (Nair et al. 2009).
Previous research has demonstrated that agroforestry system using a silvopasture approach, can
also be implemented to provide similar benefits than just a pasture system (Howlett, MosqueraLosada, et al. 2011). In addition, silvopasture has the potential of improving biodiversity, forage
productivity (quantity and quality) for animals (Jose and Dollinger 2019). Silvopasture is also
considered as one of the effective strategies to reduce atmospheric CO2 by storing carbon above
and mostly below-ground, and accumulating high net carbon from the atmosphere via
photosynthesis beneficial to the environment (4.38tC ha-1 yr-1) (Nair et al. 2009; Feliciano et al.
2018). Overall, more than two-thirds of carbon sequestered from the atmosphere is stored in the
soil and approximately 2100 Gt of carbon can be found in the terrestrial ecosystem. Vegetation
and tree growth play a meaningful role in facilitating the exchange of carbon between the
atmosphere and the soil through photosynthesis (De Deyn et al. 2008) (Figure 1.1). Some aspects,
such as plant respiration and biodegradation, enable the transfer of carbon back to the atmosphere.
However, the amount of carbon sequestered through photosynthesis and decomposition is
relatively larger; and can be a significant addition to the soil carbon pool (Bhattarai et al. 2017).

5

Figure 1.1 “Soil carbon in and output by plants and associated soil heterotrophs. Solid lines
indicate carbon incorporation and dotted lines are soil carbon loss; SOC: soil organic carbon,
VOC: volatile organic carbon, DOC: dissolved organic carbon” (De Deyn et al. 2008).
Remote sensing technology in silvopasture systems
As explained above, silvopastoral practices are normally applied to increase livestock
productivity by enhancing the quality and quantity of forage throughout the growing season (Jose
and Dollinger 2019); and to keep providing environmental services including sequestrating
atmospheric carbon above and below ground (Andrade et al. 2008). Physical on-the-ground fieldbased methods are successful techniques that have long been used to monitor silvopasture systems.
However, there is an opportunity to develop aerial remote sensing methods for a more improved,
flexible and fast quantitative method to complement physical measurement in monitoring livestock
forage production and ecosystem services (Viljanen et al. 2018). Several studies have documented
the use of remote sensing methods as alternatives to monitor forest products (Larrinaga and
Brotons 2019) and forage in grazing lands (Michez et al. 2019).
Remote sensing methods (including photogrammetry, multispectral and hyperspectral imaging)
are all processes of measuring or obtaining information of an object or an area non-invasively by
utilizing a specialized recording device without coming in direct contact with the object or area of

6
interest (Sugiura et al. 2005; Viljanen et al. 2018). Spectral remote sensing technology, imaging
spectroscopy specifically (measuring the intensity of electromagnetic radiation between a radiation
source, typically the sun and the object) has been widely applied in the biological world since the
late 1980s after launching airborne sensors and later in the 1990s with the addition of spaceborne
sensors (Pu 2017).
This spectral method has been successfully integrated recently with advanced technology such
as UAVs (Unmanned Aerial Vehicles) also known as RPAS (Remotely Piloted Aircraft system).
These tools have become more accessible to rangeland managers to efficiently monitor forage
productivity and quality for grazing animals (Michez et al. 2019). They are used also to quantify
forest aboveground biomass; by providing relevant 3-dimensional information over a larger scale
created by spectral information (Michez et al. 2019). The spectral information most commonly
used for forest and silvopasture management is provided by a combination of multiple spectral
bands in the electromagnetic spectrum including: Red, NIR (Near Infrared) and RedEdge; and
used to produce vegetation indices algorithms such as NDVI (Normalized Difference Vegetation
Index) and NDRE (Normalized Difference RedEgde) defined as:

NDVI = (NIR - Red) / (NIR + Red)
NDRE = (NIR - RedEdge) / (NIR + RedEdge)
Both indices are normally used as an indication of plant productivity including: yield assessment
(Zhang et al. 2019) as well as vegetation stress detection, plant vigor, nitrogen uptake, fertilizer
demand and leaf chlorophyll and nutrient content (Kanke et al. 2016). NDVI is derived from a
normalized transformation of the near-infrared (NIR) band to Red in the visual spectrum resulting
in a reflectance ratio. The index defines values from -1 to +1, where negative values mean the
absence of vegetation (Pettorelli et al. 2005). NDRE index is similar to NDVI but uses the ratio of
near-infrared and red-edge. The red band is replaced by red-edge band which sits between both
Red and Near Infrared (Figure 1.2), which has shown to have a lot of utility when used in remote
sensing studies with plants (Fitzgerald et al. 2006).

7

Figure 1.2: An example of a typical plant reflectance curve detailing the wavelengths and position
of visible and non-visible light (Micasense Inc.) (Parker 2019).
Research objectives
This study explored an opportunity of integrating the forest and ranching industries, to enhance
both forestry and grazing practices, so that forest production and understory forage productivity
can be fully realized. The major outcome of this research is development of improved grazing and
range management strategies through a silvopasture system by optimizing forage production where
grazing potential has been lost due to tree-canopy closure, while maintaining the environmental
sustainability. The above outcome was achieved by conducting two seasons of field experiments
and laboratory data analysis.
This thesis is separated into two data chapters focussed on the following specific objectives:
✓ Chapter 2. Test the effect of integrating forage and timber management through strip
thinning on forage productivity and plant community composition following agronomic
seed addition. Statistical comparisons of understorey vegetation were evaluated to
determine which strip thinning width has the largest influence on forage quality and
quantity, plant richness and diversity.

8
✓ Chapter 3. Assess the impact of strip-thinned areas and the addition of agronomic plant
species on soil carbon sequestration as part of climate change mitigation. The impact of
thinning on soil pH, bulk density, soil organic matter, total carbon, total nitrogen and net
carbon exchange between the soil and the atmosphere at 10 m, 15 m and 20 m widths was
evaluated statistically and compared to un-thinned control treatments.
Furthermore, remote sensing technology using a high-resolution multispectral sensor was an
additional component in the second chapter; and used as a comparative model on forage
productivity between the strip thinned areas. The results of this thesis will assist foresters and
ranchers to continue to work together to optimize their operations on a shared land base and keep
fostering ecosystem integrity including greenhouse gases reduction.

9
LITERATURE CITED
Andrade, H.J., Brook, R., Ibrahim, M., 2008. Growth, production and carbon sequestration of
silvopastoral systems with native timber species in the dry lowlands of Costa Rica. Plant
Soil 308, 11–22. doi:10.1007/s11104-008-9600-x
Baah-Acheamfour, M., Chang, S.X., Carlyle, C.N., Bork, E.W., 2015. Carbon pool size and
stability are affected by trees and grassland cover types within agroforestry systems of
western Canada. Agric. Ecosyst. Environ. 213, 105–113. doi:10.1016/j.agee.2015.07.016
Bhattarai, M.D., Secchi, S., Schoof, J., 2017. An Analysis of the Climate Change Mitigation
Potential through Soil Organic Carbon Sequestration in a Corn Belt Watershed. Environ.
Manage. 59, 77–86. doi:10.1007/s00267-016-0771-6
De Bruijn, S.L., Bork, E.W., 2006. Biological control of Canada thistle in temperate pastures
using high density rotational cattle grazing. Biol. Control 36, 305–315.
doi:10.1016/j.biocontrol.2005.10.007
De Deyn, G.B., Cornelissen, J.H.C., Bardgett, R.D., 2008. Plant functional traits and soil carbon
sequestration in contrasting biomes. Ecol. Lett. 11, 516–531. doi:10.1111/j.14610248.2008.01164.x
Delcurto, T., Porath, M., Parsons, C.T., Morrison, J.A., 2005. Management Strategies for
Sustainable Beef Cattle Grazing on Forested Rangelands in the Pacific Northwest. Soc.
Range Manag. 58, 119–127.
Feliciano, D., Ledo, A., Hillier, J., Nayak, D.R., 2018. Which agroforestry options give the
greatest soil and above ground carbon benefits in different world regions? Agric. Ecosyst.
Environ. 254, 117–129. doi:10.1016/j.agee.2017.11.032
Fike, J.H., Buergler, A.L., Burger, J.A., Kallenbach, R.L., 2004. Considerations for Establishing
and Managing Silvopastures: Forage and Grazinglands. Plant Manag. Netw. 1–12.
doi:10.1094/FG-2004-1209-01-RV.Introduction
Fitzgerald, G.J., Rodriguez, D., Christensen, L.K., Belford, R., Sadras, V.O., Clarke, T.R., 2006.

10
Spectral and thermal sensing for nitrogen and water status in rainfed and irrigated wheat
environments. Precis. Agric. 7, 233–248. doi:10.1007/s11119-006-9011-z
Frey, G.E., Fassola, H.E., Pachas, A.N., Colcombet, L., Lacorte, S.M., Renkow, M., Pe, O.,
2012. A Within-Farm Efficiency Comparison of Silvopasture Systems with Conventional
Pasture and Forestry in Northeast Argentina. Land Econ. 2–20. doi:10.3368/le.88.4.639
Hall, C., Macdiarmid, J.I., Matthews, R.B., Smith, P., Hubbard, S.F., Dawson, T.P., 2019. The
relationship between forest cover and diet quality: a case study of rural southern Malawi.
Food Secur. 11, 635–650. doi:10.1007/s12571-019-00923-0
Hao, J., Dickhoefer, U., Lin, L., Müller, K., Glindemann, T., Schönbach, P., Schiborra, A.,
Wang, C., Susenbeth, A., 2013. Effects of rotational and continuous grazing on herbage
quality, feed intake and performance of sheep on a semi-arid grassland steppe. Arch. Anim.
Nutr. 67, 62–76. doi:10.1080/1745039X.2012.761326
Harris, J.A., Hobbs, R.J., Higgs, E., Aronson, J., 2006. Ecological restoration and global climate
change. Restor. Ecol. 14, 170–176. doi:10.1111/j.1526-100X.2006.00136.x
Hoberg, G., Peterson St-Laurent, G., Schittecatte, G., Dymond, C.C., 2016. Forest carbon
mitigation policy: A policy gap analysis for British Columbia. For. Policy Econ. 69, 73–82.
doi:10.1016/j.forpol.2016.05.005
Hogarth, N.J., Belcher, B., Campbell, B., Stacey, N., 2013. The Role of Forest-Related Income
in Household Economies and Rural Livelihoods in the Border-Region of Southern China.
World Dev. 43, 111–123. doi:10.1016/j.worlddev.2012.10.010
Howlett, D.S., Mosquera-Losada, M.R., Nair, P.K.R., Nair, V.D., Rigueiro-Rodríguez, A., 2011.
Soil Carbon Storage in Silvopastoral Systems and a Treeless Pasture in Northwestern Spain.
J. Environ. Qual. 40, 825–832. doi:10.2134/jeq2010.0145
Husak, A.L., Grado, S.C., 2002. Monetary benefits in a Southern silvopastoral system. South. J.
Appl. For. 26, 159–164. doi:10.1093/sjaf/26.3.159
Jose, S., 2009. Agroforestry for ecosystem services and environmental benefits: An overview.

11
Agrofor. Syst. 76, 1–10. doi:10.1007/s10457-009-9229-7
Jose, S., Dollinger, J., 2019. Silvopasture: a sustainable livestock production system. Agrofor.
Syst. 93, 1–9. doi:10.1007/s10457-019-00366-8
Jose, S., Gillespie, A.R., Seifert, J.R., Biehle, D.J., 2000. Defining competition vectors in a
temperate alley cropping system in the midwestern USA. Agrofor. Syst. 48, 41–59.
Kanke, Y., Tubaña, B., Dalen, M., Harrell, D., 2016. Evaluation of red and red-edge reflectancebased vegetation indices for rice biomass and grain yield prediction models in paddy fields.
Precis. Agric. 17, 507–530. doi:10.1007/s11119-016-9433-1
Karki, U., Goodman, M.S., Sladden, S.E., 2009. Nitrogen source influences on forage and soil in
young southern-pine silvopasture. Agric. Ecosyst. Environ. doi:10.1016/j.agee.2008.09.007
Klopfenstein, N.B., Carman, R.C., Clason, T.R., Rietveld, W.J., Sharrow, S.H., 1997.
Silvopasture : An Agroforestry Practice. U.S. Dep. Agric. For. Serv. Natl. Agrofor. Cent. 6,
1–6.
Krzic, M., Page, H., Newman, R.F., Broersma, K., 2004. Aspen regeneration, forage production,
and soil compaction on harvested and grazed boreal aspen stands. BC J. Ecosyst. Manag. 5,
30–38.
Kulshreshtha, S., Modongo, O., Florizone, A., 2012. Economic Impacts of Livestock production
in Canada - A Regional Multiplier Analysis, Canfax Research Services, Canadian
Cattlemen’s Association.
Larrinaga, A.R., Brotons, L., 2019. Greenness Indices from a Low-Cost UAV Imagery as Tools
for Monitoring Post-Fire Forest Recovery. Drones 3, 6. doi:10.3390/drones3010006
Lin, C., McGraw, R., George, M., Garrett, H., 1999. Shade effects on forage crops with potential
in temperate agroforestry practices. Agrofor. Syst. 44, 109–119.
doi:10.1023/A:1006205116354
Lindgren, P.M.F., Sullivan, T.P., 2014. Response of forage yield and quality to thinning and

12
fertilization of young forests : implications for silvopasture management 289, 281–289.
Michez, A., Lejeune, P., Bauwens, S., Lalaina Herinaina, A.A., Blaise, Y., Muñoz, E.C., Lebeau,
F., Bindelle, J., 2019. Mapping and monitoring of biomass and grazing in pasture with an
unmanned aerial system. Remote Sens. 11, 1–14. doi:10.3390/rs11050473
Nair, P.. R., Haile, S.G., Nair, V.D., Kumar, M., 2009. Soil carbon sequestration in tropical
agroforestry systems 1099–1111.
Nilsson, S., Schopfhauser, W., 1995. The carbon-sequestration potential of a global afforestation
program. Clim. Change 30, 267–293. doi:10.1007/BF01091928
Norris, K., 2012. Biodiversity in the context of ecosystem services: The applied need for systems
approaches. Philos. Trans. R. Soc. B Biol. Sci. 367, 191–199. doi:10.1098/rstb.2011.0176
Onteru, S.K., Ampaire, A., Rothschild, M.F., 2010. Biotechnology developments in the livestock
sector in developing countries. Biotechnol. Genet. Eng. Rev. 27, 217–228.
doi:10.1080/02648725.2010.10648151
Parker, R., 2019. Plant reflectance curve and wavelength (nm). solar irradiance, the light energy
emanating from the sun as measured from the Earth after passing through the atmosphere.
[WWW Document]. Micasense Inc. URL https://support.micasense.com/hc/enus/articles/115004292333-Why-Use-Panels-with-a-Light-Sensor-?mobile_site=true
Pettorelli, N., Vik, J.O., Mysterud, A., Gaillard, J.M., Tucker, C.J., Stenseth, N.C., 2005. Using
the satellite-derived NDVI to assess ecological responses to environmental change. Trends
Ecol. Evol. 20, 503–510. doi:10.1016/j.tree.2005.05.011
Pu, R., 2017. Hyperspectral remote sensing: Fundamentals and practices. CRC Press Taylor&
Francis Group, Boca Raton, Florida.
Sanderman, J., Reseigh, J., Wurst, M., Young, M.A., Austin, J., 2015. Impacts of rotational
grazing on soil carbon in native grass-based pastures in southern Australia. PLoS One 10,
1–16. doi:10.1371/journal.pone.0136157

13
Seely, B., Welham, C., Kimmins, H., 2002. Carbon sequestration in a boreal forest ecosystem:
Results from the ecosystem simulation model, FORECAST. For. Ecol. Manage. 169, 123–
135. doi:10.1016/S0378-1127(02)00303-1
Shrestha, R.K., Alavalapati, J.R.R., Kalmbacher, R.S., 2004. Exploring the potential for
silvopasture adoption in south-central Florida: An application of SWOT-AHP method.
Agric. Syst. 81, 185–199. doi:10.1016/j.agsy.2003.09.004
Soussana, J.-F., Soussana, J.-F., Loiseau, P., Vuichard, N., Ceschia, E., Balesdent, J., Chevallier,
T., Arrouays, D., 2004. Carbon cycling and sequestration opportunities in temperate
grasslands. Soil Use Manag. 20, 219–230. doi:10.1079/sum2003234
Sugiura, R., Noguchi, N., Ishii, K., 2005. Remote-sensing technology for vegetation monitoring
using an unmanned helicopter. Biosyst. Eng. 90, 369–379.
doi:10.1016/j.biosystemseng.2004.12.011
Thornton, P.K., 2010. Livestock production: Recent trends, future prospects. Philos. Trans. R.
Soc. B Biol. Sci. 365, 2853–2867. doi:10.1098/rstb.2010.0134
Verma, P., Bijalwan, A., Dobriyal, M.J.R., Swamy, S.L., Thakur, T.K., 2017. A paradigm shift
in agroforestry practices in Uttar Pradesh 112, 509–516.
Viljanen, N., Honkavaara, E., Näsi, R., Hakala, T., Niemeläinen, O., Kaivosoja, J., 2018. A
novel machine learning method for estimating biomass of grass swards using a
photogrammetric canopy height model, images and vegetation indices captured by a drone.
Agric. 8. doi:10.3390/agriculture8050070
Wang, J., Li, Y., Bork, E.W., Richter, G.M., Eum, H. Il, Chen, C., Shah, S.H.H., Mezbahuddin,
S., 2020. Modelling spatio-temporal patterns of soil carbon and greenhouse gas emissions in
grazing lands: Current status and prospects. Sci. Total Environ. 739, 139092.
doi:10.1016/j.scitotenv.2020.139092
Willms, W., Bailey, A.W., McLean, A., 1980. Effect of Burning or Clipping Agropyron
spicatum in the Autumn on the Spring Foraging Behaviour of Mule Deer and Cattle. J.
Appl. Ecol. 17, 69. doi:10.2307/2402964

14
Wilson, M.H., Lovell, S.T., 2016. Agroforestry-The next step in sustainable and resilient
agriculture. J. Sustain. 8, 1–15. doi:10.3390/su8060574
Zampaligré, N., Schlecht, E., 2018. Livestock foraging behaviour on different land use classes
along the semi-arid to sub-humid agro-ecological gradient in West Africa. Environ. Dev.
Sustain. 20, 731–748. doi:10.1007/s10668-017-9907-y
Zhang, K., Ge, X., Shen, P., Li, W., Liu, X., Cao, Q., Zhu, Y., Cao, W., Tian, Y., 2019.
Predicting rice grain yield based on dynamic changes in vegetation indexes during early to
mid-growth stages. Remote Sens. 11. doi:10.3390/rs11040387

15

CHAPTER 2 – UNDERSTORY PLANT COMMUNITY ESTABLISHMENT
WITHIN A LODGEPOLE PINE SILVOPASTURE SYSTEM
INTRODUCTION
British Columbia (BC) and elsewhere around the world have been investing in forest
management and conservation practices for biodiversity (Sullivan et al. 2001) including lodgepole
pine (Pinus contorta) species (Woods 2003). Pinus contorta is among the dominant coniferous
tree species in BC, representing 41% of total tree seedlings planted, with high growth success in a
wide range of forest site conditions (Woods et al. 2000). This species occupies 14.9 million
hectares of BC forestland. Although Lodgepole pine is described as a low shade tolerance species
(Axelson et al. 2009), once fully established can affect the growth of understory plant species.
Modern forest management is increasingly considering forest biodiversity in planning, including
plant community growth which is considered a major food source for wildlife and livestock
(Thomas et al. 1999). Therefore, the density of tree crown closure is an important consideration in
managing forests (Zaborske et al. 2002) since understory vegetation depends highly on the amount
of sunlight, nutrients and soil moisture content (Lindgren and Sullivan 2014). Dense crown closure
and soil nutrients are major factors that affect the timber and forage supply in BC, as well as
ecological disturbances such as forest fire and mountain pine beetle (Dendroctonus ponderosae
Hopkins) (Axelson et al. 2009). The mountain pine beetle is a North American native forest insect
that attacks and kills forest tree species. Lodgepole pine stands have suffered the greatest beetle
outbreak, causing mortality in up to 20 million ha of pine forests in Canada and the United states
(Alfaro et al. 2015).
Silvopasture practice such as thinning can be an effective approach to diversify forest services
including understory diversity and composition (Thomas et al. 1999). Forest thinning, a
silvicultural practice that selectively removes trees to facilitate healthy growth of remaining trees
and understory vegetation (Verschuyl et al. 2011), has many benefits on forage production in a
forested area; such as: increasing forest floor light by reducing canopy closure, reducing soil
nutrients competition, and improving soil moisture availability (Pang et al. 2013). McConnell and
Smith (1970) studied the response of understory vegetation to thinning versus un-thinned areas,
and found that thinned area increased grasses by 51%, forbs by 37% and shrubs by 12% with a
significant increase in timber production. In addition, thinning can also be useful to reduce wildfire
severity (Verkaik and Espelta 2006) and increase soil temperature due to increased solar insolation

16
on the ground surface, which can have a positive effect on the performance of soil microbial
activities; and facilitates the conversion of organic N to a plant available form (Pang et al. 2013).
In general, forage productivity depends heavily on a healthy soil and the capacity of microbial
activity in the soil. Temperature plays a significant role in influencing soil microorganisms’
activities including organic matter decomposition and recycling plant material (Pietikäinen et al.
2005). Some soil microorganisms (bacteria and fungi) associate symbiotically with plant roots; for
instance, white clover (Trifolium repens L.) for the fixing of nitrogen, which is an essential element
for plant growth (Caradus et al. 1996). Lindgren and Sullivan (2014) examined the potential
influence of thinning on understory vegetation quality and quantity in young forests; and found
that thinning influences vegetation growth by improving the amount, distribution and timing of
forage use. However, little is known on the appropriate thinning distance and target canopy density
for enhancing forage yield and quality, particularly with respect to the many different tree species
used in silvicultural practices.
This chapter aims to test various widths of strip thinning, a form of forest stand thinning, in a
lodgepole pine forest for improving forage yield and the factors that might influence nutritional
value and digestibility of forage for livestock feed such as Crude Protein (CP), soluble
carbohydrate, fat, lignin, Total Digestible Nutrients (TDN), Acid Detergent Fiber (ADF) and
Neutral detergent Fiber (NDF) as a response to strip thinning. A balance of nutrients and digestible
fiber is highly recommended in order to maintain animal health as the quality depends on the ratio
of positive nutrients (CP, carbohydrates, fat, and minerals) and negative nutrients (NDF, ADF, and
lignin) which are the most used parameters to determine forage nutritional value (Uniyal et al.
2005; Zeng and Chen 2018). The addition of high quality and leafy, palatable forage species can
be a good complement to native species in increasing nutritional value in a grazing area while
maintaining a diverse, sustainable ecosystem (Svejcar and Vavra 1985).
Schweitzer et al (1993) showed that over-seeding nutritious agronomic forage species (nonnative species) can be a beneficial addition to native species in optimizing beef production, and
enhancing wildlife habitat while maintaining timber yields. However, some non-native species,
once fully established in an area, can pose a significant threat to native ecosystems. Non-native
species have the ability to dominate and alter native ecosystems. Some of the non-native plant
species categorized as invasive species can eliminate less competitive neighbouring plants by
releasing compounds that may modify soil chemistry (Weidenhamer and Callaway 2010).

17
However, there are some species which are socially, economically and environmentally beneficial
to the ecosystem by providing desirable ecosystem functions (Schlaepfer et al. 2011) such as
orchardgrass (Dactylis glomerate), white clover (Trifolium repens) (Orefice et al. 2019) and
intermediate wheatgrass (Thinopyrum intermedium) (Wills et al. 1998).
The increase of capability in scientific research has led to a success in developing more
advanced methods to manage ecosystem resources including: in depth assessment of forage yield
and nutritional value (Insua et al. 2019); as well as mapping, and modeling the distribution, and
effect of invasive species on native plant species (Joshi et al. 2004). Satellite based remote sensing
technology is one of the methods currently adopted in this research project to monitor understory
vegetation growth and health through photosynthetic activities (Michez et al. 2019). Despite the
recent high popularity of remote sensing in assisting natural resource managers and foresters, the
use of remote sensing technology has been recommended for almost four decades (Tueller 1989)
in estimating: above ground vegetative biomass; forage quality such as crude protein, different
detergent fibers and soluble carbohydrate in the field (Zeng and Chen 2018), and also measuring
and monitoring soil carbon sequestration (Goetz and Dubayah 2011). Multispectral remote sensing
application has greatly attracted the attention of several scientists worldwide in extracting plant
phenotypic information including monitoring plant growth and development (Deng et al. 2018).
The NDVI (Normalized Difference Vegetation Index), derived from spectral remote sensing
applications, is among several indices commonly used in ecological studies as predictors for forage
productivity parameters such as: net primary production, active radiation for plant photosynthetic
rate, leaf area index, evapotranspiration and plant biomass (Borowik et al. 2013).
This chapter summarizes the results from the field experiment conducted before and after strip
thinning lodgepole pine forests in the southern interior of British Columbia, Canada. Data were
collected using physical ground data collection and remote sensing data collection using a
multispectral sensor onboard a RPAs with the objectives of: 1) evaluating plant community
growth, species richness and diversity of both native plant species and agronomic species (seeded)
as a response to strip- thinning; 2) assessing the effect of strip- thinning on palatable forage quality,
health, and utilization; and, 3) use of NDVI and NDRE index products derived from remote
sensing to assess forage growth and availability.

18
The overall goal of this study was to integrate ranching and forest industries by increasing the
productivity of feed in terms of volume and quality on the same land-base. This approach will
contribute in optimizing land use through a silvopasture model created by strip thinning.

MATERIALS & METHODS
Site Description
To evaluate the integration of forest and grazing management approaches, as well as continuous
monitoring of soil carbon sequestration, an operational-scale pilot was established in a midrotation forest stands of 45-year old lodgepole pine. The stands were harvested at the end of July
2018 at 10 m, 15 m and 20 m width strips across 101.4 hectares of the three adjacent forest sites,
situated in Goudie, Kelowna, British Columbia at an elevation range between 1340 – 1400 m. The
10 m width strips were separated by 20 m timber buffers, and 15 m and 20 m width strips were
separated by 30 m and 40 m timber buffers respectively (Appendix-A.15). The location of the first
block (32.8 ha) was at 49°55'24"N latitude and 119°14'37"W longitude with a minimum elevation
of 1340 m and a maximum elevation of 1380 m. The second block (42.2 ha) was located at
49°56'9"N latitude and 119°14'25"W longitude with a minimum elevation of 1380 and maximum
elevation of 1400 m and the third (26.4 ha) at 49°56'30"N latitude and 119° 14'50"W longitude
with a minimum elevation of 1380 m and 1400 m and a general average slope of 11.3% (Arithmetic
slope). Strip thinning harvest method was implemented following a randomized complete block
design with four treatments per block (Figure 2.1). On average, nine strips were designed in 10 m
width treatments (average length of 297 m), seven strips in 15 m width treatments (average length
of 256 m) and five strips in 20 m width treatments (average length of 241 m) in each of the three
adjacent blocks (Figure 2.1).
For data collection purposes, two strips were selected in 10 m, 15 m, and 20 m strips including
two plots in the uncut control areas and sample data were collected from the center areas of the
strip toward north for strips facing north and east for strips facing east.

19

Figure 2.1: An operational- scale pilot map initiated at a 101.4 ha area in the Okanagan region,
Goudie, Kelowna, BC. The experiment was conducted on three blocks with four treatments. The
red marked areas are control treatments. The green strips are 10 m wide, purple strips are 15 m
wide and blue strips 20 m width. The yellow pins indicate the sampling points.
Research Design
Four highly palatable agronomic seed mix species were selected and seeded at 12 kg/ha in
October 2018 after timber harvesting (Table 2.1). Species selected were based on the fact that they
are preferred by livestock and wild animals, and suitable to intensive rotational grazing systems
(USDA- United States Department of Agriculture 2019).

20
Table 2.1. Highly palatable agronomic plant species used for improving the quantity and
quality of animal forage

October, 2018

Ratio
(%)

Additional seed mix
1 Dactylis glomerata

Orchard grass

30%

2 Bromus riparius

Meadow brome

30%

3 Thinopyrum intermedium

Intermediate wheatgrass

30%

4 Trifolium repens

White clover

10%

These four forage species are highly palatable to all classes of livestock and wildlife and are
one of the earliest species to initiate growth in the spring with highly significant growth during
cool conditions (USDA- United States Department of Agriculture 2019). They are very resistant
to winter conditions but less productive under extreme hot conditions, saline soils and wet or
poorly drained areas (USDA- United States Department of Agriculture 2019).
Pre-harvest field data collection
Baseline data including a percent cover understory vegetation survey were collected before strip
thinning activities in June 2018. A 50 m transect was established in the middle of four designated
treatment plans. Five sampling points per selected strip were established at 10 m intervals along
the 50 m transect. At each sampling interval, a 0.5 m by 0.5 m quadrat was placed, and percent
cover data by species were recorded in order to estimate plant species composition, richness and
diversity (Figure 2.2) (Daubenmire 1984).

21

Figure 2.2. An extended Daubenmire method of a 0.25 sq. m quadrat used to collect % cover data
per species within three adjacent forest sites located in Goudie, Kelowna, British Columbia before
strip harvesting activities.
Post-harvest field data collection.
Post-harvest data collection was carried out in summer 2019 and 2020 after timber harvesting
and seeding in the designed blocks/areas. The strip thinning harvest method provided an
opportunity for additional forage by reducing forest canopy cover, nutrients competition while
retaining the same percentage of timber as a conventional thinning approach. Field data collection
and remote sensing were used in the designated treatments to explore the effects of strip harvesting
on plant productivity, plant community composition, richness, and diversity as well as evaluate
the long-term suitability of using additional agronomic plant species known to increase Animal
Unit Months within the widths of the designed strips.
The Daubenmire cover class method was carried out in July, August 2019 and July 2020 to
visually assess percent cover by species and productivity within quadrats (Bonham et al. 2004). A
50 m transect was established from the designated sampling points parallel to the thinned strips
(Figure 2.1). A total of 10 sampling points (using 0.5 m by 0.5 m quadrats) were recorded in each
treatment unit (5 sampling points per strip) to provide the estimates of plant species composition,
species richness and diversity post timber harvest. Biomass was clipped within each quadrat at
ground level without considering the previous year’s litter. Samples were separated into seeded
species, unseeded-native palatable and unseeded non-palatable species (Appendix. A.16) (USDAUnited States Department of Agriculture 2019). All clipped samples were dried at 70oC for 48h in

22
a forced convection constant temperature drying oven, DKN 812 series of Yamato scientific Co.,
Ltd, then weighed on an analytical scale to calculate biomass yield in gram per m2 (Pavlů et al.
2006). Seeded and unseeded-native palatable species were ground to pass a 1 mm screen and
analyzed using a FOSS high-performance InfraXactTM based NIR and transflectance analyzer with
a scanning range of 570 - 1850 nm (FOSS Analytical 2010) in order to determine the essential
parameters usually used to assess energy and digestibility level in forage such as CP (Crude
Protein), S.Carb (Soluble Carbohydrates), Fat, Lignin, ADF (Acid Detergent Fiber), NDF (Neutral
Detergent Fiber) and TDN (Total Digestible Nutrients) (Zeng and Chen 2018). NIR spectrum used
by FOSS InfraXact is above the visible and Middle InfraRed (MIR) region of the electromagnetic
spectrum.
Remote sensing data collection
An improved DJI Matrice 210 RTK V2 quadcopter was equipped with a Micasense Altum
camera (Figure 2.3) for monthly aerial flights over the study area during the 2019 and 2020
growing seasons. The DJI Matrice 210 RTK (Real-Time Kinematic) Version 2 is an improved
high precision quadcopter equipped with an upgraded positioning system mobile ground station
consisting of a high precision GNSS (Global Navigation Satellite System) receiver (Figure 2.4).
The Micasense Altum camera is a multispectral camera with five separate high-resolution bands
including: Blue, Green, Red, RedEdge, and NIR with an addition of a radiometric thermal camera
and a sun irradiance sensor (DLS 2 contains an integrated GPS) (Figure 2.3) to measure specific
incoming solar irradiance for radiometric correction. The Altum sensor resolution has a GSD
(Ground Sample Distance) of 5.2 cm per pixel at 120 m AGL (Above Ground level) and 81cm per
pixel for thermal camera at 120 m.
The canopy reflectance data were collected on a clear-sky day twice during vegetation peak
productivity from late July 2019 to August 2019 and repeated in July 2020. The drone was
deployed perpendicular to the designed thinned strips at a flying height of 70 m AGL (Above
Ground Level) and a flying speed of 3 ms-2 in order to cover the entire area of 101.4 ha (Fig. 2.1)
and targeted vegetation growth in the open strips as suggested by Micasense, Inc. Prior to spectral
reflectance measurement over the study area, the DLS 2 was mounted on the drone and images of
a calibrated white reflectance panel (Figure 2.4) were recorded before and after each flight to
calibrate raw pixel to absolute reflectance images in order to provide more accurate and reliable
data. The flight missions followed the same flight plan with a front overlap and side overlap of

23
75% as recommended by Micasense, Inc. To ensure accuracy of Altum images geo-records, the
RTK mobile ground station was paired to the drone RTK receivers, and we ensured a maximum
connection was provided throughout the entire flight missions.

Figure 2.3. Drone (UAV currently known as RPA) equipped with a high resolution multispectral
camera with a blue (475 nm center, 20 nm bandwidth), green (560 nm center, 20 nm bandwidth),
red (668 nm center, 10 nm bandwidth), red edge (717 nm center, 10 nm bandwidth), near-IR (840
nm center, 40 nm bandwidth) narrow bands and thermal sensor (LWIR: 8- 14μm).

Figure 2.4. D-RTK 2 mobile station (left), a high-precision global navigation satellite system
receiver compatible with a new aircraft version, it uses navigational satellites from other networks
in providing Aircraft accuracy. Multispectral sensor calibrated reflectance panel (right).

24
Remote sensing data processing
Professional photogrammetry software (PiX4D Mapper Pro- Educational version 3.1.23) was
used for processing multispectral images using a standard AfM (Structure from Motion)
(Barbasiewicz et al. 2018). Calibration images recorded prior to each mission deployment were
automatically added to the software processing memory. The software assessed each image
geolocation, computed manual tie point positions, and image overlap in the first step of data
processing. The second step consisted of constructing a point cloud and mesh and, the final step
of the process consisted of building a DSM (Digital Surface Model), Orthomosaic and Index
products including the normalized difference vegetation index (NDVI) and normalized difference
red edge (NDRE) used to assess forage availability and productivity (Zhang et al. 2019). As
detailed in Chapter 1, both index products were built from red, NIR and red_edge bands processed
in the final step of software data processing.
The mean NDVI and NDRE values were retrieved using an advanced ArcGIS Desktop version
10.7.0.10450 and the images in TIFF format from the processed index products layers were
uploaded into ArcMap workflow. The GPS information recorded during field data collection,
which consists of strip location, sampling point locations, and distance from the sampling plots to
the buffer zone, were used to mark the sampling points. Five-polygon shapefiles (approx. 0.5 m
by 0.5 m quadrats) were created at each 10 m along the 50 m line (transect) drew from the pinned
points following the WGS 1984 UTM Zone 11N coordinate system. The polygonal shapefile of
each plot was used to extract digital values from NDVI and NDRE layers; and the mean values for
each sampling point were calculated by Zonal statistic in ArcGIS.

STATISTICAL ANALYSIS
Data analysis and graphical outputs were completed using R version 3.6.2 (The R Foundation
for Statistical Computing). Vegetation weight per treatment, species richness, diversity as well as
forage nutrient parameters, and remote sensing data were tested for normality using the shapiro
test (Jensen 2009; Mohd Razali and Bee Wah 2011). Variances within groups were tested for
homogeneity using the Fligner-Killeen test (Conover et al. 1981) and when necessary, data were
transformed using a natural logarithm or a square root function (sqrt). Shannon Wiener Diversity
Index (H’) was used to assess understory species diversity and count species richness of each
quadrat in the four treatments.

25
The designed experiment allowed a comparison of means among the four treatments in the three
adjacent blocks design using an analysis of variance test for forage productivity as well as NDVI
and NDRE analysis. Analysis of variance test was followed by the Tukey post-hoc test to find
treatments that were significantly different from each other at a 5% probability level. Kruskal Wallis
tests followed by Dunn’s post-hoc test, using the Benjamini-Hochberg procedure to control for
multiple comparisons were applied to the data that did not follow a normal distribution and equal
variance assumptions. Analysis of covariance (ANCOVA) was employed to control for a priori effect
pre-timber harvest 2018.

A Principal Coordinates Analysis (PCoA) using Bray Curtis distance matrix was a good
ordination method applied to visualize and compare understory species community composition
in each treatment unit for data collected pre-harvest 2018. A Jaccard distance matrix was applied
to data collected post-harvest 2019 and 2020. PCoA and Jaccard similarity index were able to
converge, capture and explain more variance in our post-harvest 2019 and 2020 dataset; and Bray
Curtis captured more variance explained in species surveyed in July 2018. PCoA is theoretically
an extension of Principal Component Analysis (PCA), and widely applied in ecology and used
with any dissimilarity matrix (Paliy and Shankar 2016).
Permutational multivariate analysis of variance (PERMANOVA) was performed on the same
PCoA matrix (Number of permutations = 999) and tested if understorey species communities were
similar among the different treatment units (Control, 10m, 15m, 20m strip). PERMANOVA is a
semiparametric method, and appropriate for community composition partitioning, and a powerful
tool for the analysis of similarity or variation in plant species community composition (Anderson
2017).

RESULTS
Forage productivity
Yield
After vegetation survey, biomass was harvested in July 2019, August 2019 and July 2020 in
order to estimate biomass production between treatment units. Due to a high rainy season and the
late germination of the agronomic seed mix, our analysis did not consider the data collected in July
2019 (Canada Government 2019). August 2019 biomass yield was higher in 20 m thinned
treatments, but a statistical analysis did not show any significant difference between the treatment

26
units (p >0.05). However, biomass harvested in July 2020 showed a significant effect of strip
thinning on vegetation growth, with a higher forage yield in 20 m and 15 m thinned treatments
compared to 10 m thinned treatments, and very low yield in the uncut control (p < 0.05). The uncut
control treatment was different from all the thinned treatments. The 15 m and 20 m thinned
treatments did not show any difference (Figure 2.5).

Figure 2.5 Mean total biomass weight in gram per 0.25 sq.m for understory vegetation harvested
in each treatment in August 2019 and July 2020. Error bars represent standard error of the mean
(n= 6 for each group in 2019 and 30 for each group in 2020). (NS= non-significant; * = p≤ 0.05;
** = p≤ 0.01; *** = p≤ 0.001; **** = p≤ 0.0001).
Quality
Both seeded agronomic species and unseeded-native species were harvested and analyzed
separately in August 2019 and July 2020, for determining the amount of forage quality content in
control, 10, 15 and 20 m thinned treatments. The quality analysis of agronomic species added post

27
timber harvesting excluded the uncut control treatment because they were not part of the seeded
treatment units. Seven major parameters were selected to determine forage quality content in each
treatment unit: CP (Crude Protein), Soluble Carbohydrate, fat, lignin, NDF (Neutral Detergent
Fiber), ADF (Acid Detergent Fiber) and TDN (Total Digestible Nutrients) (Asekova et al. 2016).
All the seven nutrients parameters assessed post-harvest 2019 and 2020 (Table 2.2 & table 2.3)
were compared between 10 m, 15 m and 20 m thinned treatments including uncut control treatment
for native palatable species and without uncut control treatment for seeded agronomics (Appendix.
A.16). No influence of any strip thinning on forage quality was observed at any thinning width
including the uncut controls, both in agronomic seed mix added (Table 2.2), and unseeded native
palatable species (Table 2.3) (p > 0.05).
Table 2.2: Post timber harvest 2019 and 2020 mean understory seeded agronomic
vegetation harvested in the opening strips.

TDN
CP
NDF
Lignin
FAT
ADF
Sol_Carbos
TDN
CP
NDF
Lignin
FAT
ADF
Sol_Carbos

August 2019 seeded agronomic species
10 m
15 m
20 m
59.62 ± 3.52
59.66 ± 0.49
59.66 ± 5.42
7.98 ± 1.03
8.56 ± 0.26
6.66 ± 0.94
52.42 ± 2.88
54.27 ± 1.35
50.73 ± 4.18
8.01 ± 0.98
9.37 ± 1.68
7.18 ± 0.90
2.39 ± 0.14
2.95 ± 0.70
2.26 ± 0.16
38.35 ± 3.16
38.3 ± 0.44
38.32 ± 4.87
9.18 ± 0.67
9.72 ± 0.71
9.45 ± 0.71
July 2020 seeded agronomic species
53.49 ± 3.03
54.27 ± 3.22
56.78 ± 2.55
9.92 ± 1.24
9.40 ± 1.49
10.25 ± 0.83
55.07 ± 1.45
54.90 ± 0.80
53.44 ± 1.25
6.74 ± 0.21
6.31 ± 0.19
6.41 ± 0.29
4.19 ± 1.14
4.10 ± 1.35
2.94 ± 0.53
43.86 ± 2.72
43.16 ± 2.89
40.91 ± 2.29
11.72 ± 1.82
15.84 ± 1.62
13.63 ± 1.15

P-Value
1
0.37
0.76
0.56
0.49
1
0.9
0.58
0.82
0.59
0.36
0.37
0.58
0.11

Values are means ± standard error in august 2019 (n= 3) and July 2020 (n= 6). TDN= Total Digestible Nutrients.
CP= Crude Protein. NDF= Neutral Detergent Fiber. ADF= Acid Detergent Fiber. Treatments means were
compared using ANOVA test (significant level p= 0.05).

28
Table 2.3: Post timber harvest 2019 and 2020 mean understory native vegetation harvested
in all treatment units.
August 2019 Unseeded native palatable species
10 m
15 m
20 m
Uncut control P-value
TDN
58.79 ± 2.82
62.94 ± 1.69 59.58 ± 3.56
56.94 ± 0.73
0.51
CP
8.81 ± 0.77
11.07 ± 1.10 7.15 ± 1.07
8.68 ± 0.25
0.071
NDF
62.93 ± 4.07
53.60 ± 4.34 61.52 ± 6.17
63.34 ± 2.87
0.4
Lignin
7.17 ± 0.85
7.15 ± 0.07
6.21 ± 0.47
7.27 ± 0.10
0.55
FAT
2.84 ± 0.62
2.48 ± 0.61
2.87 ± 0.35
2.16 ± 0.46
0.77
ADF
39.1 ± 2.54
35.38 ± 1.52 38.39 ± 3.20
40.77 ± 0.66
0.51
Sol_Carbos 8.92 ± 1.33
9.11 ± 0.50
8.53 ± 1.12
6.54 ± 0.25
0.36

July 2020 Unseeded native palatable species
10 m
15 m
20 m
Uncut control P-value
TDN
53.58 ± 0.96
53.50 ± 1.04 52.94 ± 2.57
51.80 ± 3.80
0.95
CP
9.58 ± 0.88
9.23 ± 0.92
11.01 ± 1.41
12.55 ± 1.21
0.19
NDF
61.70 ± 4.41
60.87 ± 3.22 58.05 ± 2.20
57.30 ± 1.95
0.67
Lignin
7.26 ± 0.44
6.83 ± 0.13
6.52 ± 0.20
6.94 ± 0.19
0.22
FAT
3.46 ± 0.78
3.39 ± 0.54
4.74 ± 0.95
5.40 ± 1.33
0.38
ADF
43.78 ± 0.86
43.85 ± 0.93 44.36 ± 2.31
45.38 ± 3.42
0.95
Sol_Carbos 7.95 ± 1.52
9.73 ± 1.01
13.21 ± 0.91
10.8 ± 1.80
0.09
Values are means ± standard error in august 2019 (n= 3) and July 2020 (n= 6). TDN= Total Digestible Nutrients.
CP= Crude Protein. NDF= Neutral Detergent Fiber. ADF= Acid Detergent Fiber. Treatments means were
compared using ANOVA test (significant level p= 0.05).

Species richness, and diversity
Percent cover per species identified and collected in June 2018, August 2019 and July 2020
were used to determine species richness and diversity. Understory plant species richness and
diversity pre-harvest June 2018 was significantly different between 15 m and 10 m thinned
treatments as well as uncut control treatments (p < 0.05). A high species richness was found in the
areas assigned to the 15 m treatments, and low in the areas assigned to the uncut control, and 10
m treatments (Figure 2.6). Species diversity was higher in the areas assigned to the 15 and 20 m
strip treatments, and lower in the areas assigned to the 10 m strip treatments (p< 0.05) (Figure 2.7).

29
Understory species richness and diversity assessed post-harvest August 2019 were not affected
by any strip-thinned width. In addition, we did not observe any effect in uncut control treatment
as well (p > 0.05) (Figure 2.6& 2.7). However, Post harvest July 2020 found a significant
difference in species richness (Figure 2.6) and diversity (Figure 2.7) between 15 m, 20 m thinned
treatments and uncut control treatment (p < 0.05), with a higher species richness and diversity in
the 15 m and 20 m thinned strips and very low species richness and diversity in the uncut control
treatment units (Figure 2.6 & 2.7).
The analysis of covariance (ANCOVA) was applied to test the interaction effects between preharvest 2018 and post-harvest 2019 and 2020. The results obtained found a significant difference
(p < 0.05), with a high species richness and diversity in 20 m and 15 m thinned treatments and low
in uncut control. 15 m was also significantly higher than 10 m thinned treatments (Table 2.4).

Figure 2.6 Mean species richness for understory vegetation harvested in each treatment unit preharvest 2018, post-harvest 2019 and 2020. Error bars represent standard error of the mean (n=
30 for each group). (NS= non-significant; * = p≤ 0.05; ** = p≤ 0.01; *** = p≤ 0.001; **** = p≤
0.0001).

30

Figure 2.7 Shannon-Wiener (H) mean species diversity for understory vegetation harvested in
each treatment unit pre-harvest 2018, post-harvest 2019 and post-harvest 2020. Error bars
represent standard error of the mean (n= 30 for each group). (NS= non-significant; * = p≤ 0.05;
** = p≤ 0.01; *** = p≤ 0.001; **** = p≤ 0.0001).

31
Table 2.4: ANCOVA and post-hoc test results for understory species richness and diversity
pre-harvest 2018 and post-harvest 2019-2020.

15 m - 10 m
20 m - 10 m
Control - 10 m
20 m - 15 m
Control - 15 m
Control - 20 m

15 m - 10 m
20 m - 10 m
Control - 10 m
20 m - 15 m
Control - 15 m
Control - 20 m

Species Richness
Estimate Std. Error
1.1111
0.3612
0.5444
0.3612
-0.6111
0.3612
-0.5667
0.3612
-1.7222
0.3612
-1.1556
0.3612
Species diversity
Estimate Std. Error
0.20227 0.05702
0.13221 0.05702
-0.01734 0.05702
-0.07006 0.05702
-0.21961 0.05702
-0.14955 0.05702

t value
3.077
1.508
-1.692
-1.569
-4.769
-3.2

Pr(>|t|)
0.01191 *
0.43406
0.32939
0.3977
< 0.001 ***
0.00803 **

t value
3.547
2.319
-0.304
-1.229
-3.851
-2.623

Pr(>|t|)
0.00274 **
0.09552.
0.99022
0.60902
< 0.001 ***
0.04482 *

Understory community composition
Understory species community composition was identified individually in each treatment unit
within each of the three blocks. In June 2018, thirty-eight species were identified but only five
species were reasonably palatable to livestock with an extremely low percent cover (Table 2.5).
Pine grass (Calamagrostis rubescens) was a dominant grass species and covered 5% of the entire
area but was less dominant than some shrubs species such as grouseberry (Vaccinium scoparium)
(8% cover of the entire area).

32
Table 2.5. Estimate percent cover of six native grazed plant species on three experimental
design blocks pre-harvest 2018 in Goudie, Kelowna B.C
Scientific name
Calamagrostis rubescens

Common name
Pinegrass

% cover
5

Carex spp.

Sedges

0.3

Osmorhiza berteroi

Mountain sweet cicely

1.1

Chamerion angustifolium, also
recognised as Epilobium
angustifolium

Fireweed

0.2

Taraxacum erythrospermum

Dandelion

0.1

PERMANOVA and PCoA were used to compare and plot treatment units for the year 2018.
Variation in species community composition was seen across the treatment units at a p < 0.01
(Figure 2.8). By comparing each treatment individually, PERMANOVA showed a significant
variation between uncut control and the areas designated to be thinned at 15 m width (p < 0.05)
(Figure 2.10). The relationship between the understory species composition of the uncut control
community and the areas designated to be thinned at 10 m and 20 m width was not significant (p
> 0.05) (Figure 2.9 & 2.11). Bray-Curtis distance matrix measured treatment units’ species
community composition, and the first PCoA explained 15.85 percent of the total variance, while
the second PCoA explained 14.14 percent of the variation in community composition between
uncut control and the areas assigned to be 10 m, 15 m, and 20 m treatments. PCoA with Bray
Curtis distance matrix between control and areas assigned for the 10 m treatments explained 19.74
percent and 12.35 percent of the total variances. Uncut Control and areas assigned for 15 m
treatment were measured with the same distance matrix and PCoA explained 15.9 percent and 14.8
percent of the total variances. The final comparison between control and areas assigned for 20 m
treatments was explained by 18.04 percent and 13.17 percent of total variances.

33

Figure 2.8 PCoA (Principal Coordinates Analysis) ordination method for year 2018 pre-harvest
species community composition of four treatment units in the three adjacent forest areas.

Figure 2.9 PCoA for year 2018 pre-harvest species community composition comparing uncut
control and 10 m width treatments in the three adjacent forest blocks.

34

Figure 2.10 PCoA comparison between uncut control and 15 m width treatments in the three
adjacent forest blocks pre-harvest June 2018.

Figure 2.11 PCoA comparison between uncut control and 20 m width treatments in the three
adjacent forest blocks pre-harvest June 2018.

35
In August 2019, a total of fifty understory plant species were observed. Thirteen high to moderately
palatable species were identified during species survey including the seeded agronomic species
(Table 2.6). Orchardgrass (Dactylis glomerata L.) was a dominant species covering 8.6% of the
entire area.
Table 2.6. Estimate percent cover of thirteen grazed plant species on three experimental
design blocks post-harvest 2019 in Goudie, Kelowna B.C
Scientific name

Dactylis glomerata L.
Trifolium repens
Thinopyrum intermedium
Bromus riparius

Common name
Agronomic seeded species
Orchard grass
White clover
Intermediate
wheatgrass
Meadow brome

Native palatable species
Puccinellia spp.
Alkali grass
Calamagrostis rubescens
Pinegrass
Calamagrostis canadensis
Canada Blue joint
Poa pratensis L.
Kentucky Bluegrass
Carex spp.
Sedges
Festuca idahoensis Elmer
Idaho fescue
Osmorhiza berteroi
Mountain sweet cicely
Taraxacum erythrospermum Dandelion
Chamerion angustifolium
Fireweed

% cover

8.6
3.5
0.9
2.1

0.5
2.5
2.3
0.3
1
1.6
0.4
0.2
0.3

Plant community composition was significantly different across all treatment units (Control, 10 m,
15 m, and 20 m strips) (p < 0.001). Variation in community composition between 10 m, 15 m, 20
m strip thinned treatments and uncut control was represented by the first PCoA; and explained by
16.54 percent of the total variance, while the second PCoA explained 10.87 percent (Figure 2.12).
The first PCoA between the uncut control and the 10 m thinned treatments explained 13.08 percent
and the second explained 7.99 percent of the total variance. The uncut control was compared again
with 15 m thinned treatments and the first PCoA explained 13.23 percent of the total variance
while the second explained 8.36 percent. Uncut control treatment was finally compared with 20 m
thinned treatments and the analysis explained 10.31 percent and 7.93 percent of the total variance
(Figure 2.13).

36

Figure 2.12 PCoA comparing understory species community composition between Uncut control
treatment, 10 m, 15 m and 20 m strip thinned treatments one year post timber harvest (August
2019).

37

Figure 2.13 PCoA analysis comparing uncut control treatment to 10 m, 15 m and 20 m thinned
treatments individually after one-year post harvest (August 2019).

In July 2020, a total of fifty-nine understory plant species were observed. Fifteen high to
moderately palatable species were identified during species survey including the seeded agronomic
species (Table 2.7). Alkali grass (Puccinellia spp.) was a dominant species which covered 24.1%
of the entire area.

38
Table 2.7. Estimate percent cover of fifteen grazed plant species on three experimental
design blocks post-harvest 2020 in Goudie, Kelowna B.C
Scientific name

Common name

Agronomic seeded species
Orchard grass
White clover
Intermediate
Thinopyrum intermedium
wheatgrass
Bromus riparius
Meadow brome
Dactylis glomerata L.
Trifolium repens

Native palatable species
Puccinellia spp.
Alkali grass
Calamagrostis rubescens
Pinegrass
Festuca occidentalis
Western fescue
Calamagrostis canadensis
Canada Blue joint
Bromus inermis
Smooth brome
Koeleria macrantha
Prairie junegrass
Carex spp.
Sedges
Festuca idahoensis Elmer
Idaho fescue
Mountain sweet
Osmorhiza berteroi
cicely
Taraxacum erythrospermum
Dandelion
Chamerion angustifolium
Fireweed

% cover

22.3
14.1
11
8

24.1
23.5
20.7
11.6
11.7
11.7
9.5
7.4
6.6
5.6
5

Plant community composition was significantly different across all the treatments (Control, 10 m,
15 m, and 20 m strips) (p < 0.001). Community composition of all treatment units was represented
by PCoA method; and explained by 9.8% of the total variance, while the second PCoA explained
17% (Figure 2.14). A significant variation in species community composition was found between
uncut control and 10 m thinned treatment units and PCoA explained 13% and 22% of the total
variance (p < 0.001) (Figure 2.15). The comparison between uncut control and 15 m thinned
treatment units was also significantly different in species community composition (p < 0.001), and
PCoA explained 11% and 20% of the total variance (Figure 2.16). Uncut control was finally paired
with 20 m thinned treatments and both treatment units were different in species community
composition (p< 0.001), and the analysis explained 10% and 18% of the total variance (Figure
2.17). Variation across strip thinned widths was also noticed after comparing 10 m, 15 m and 20
m strip widths (p< 0.01) (Appendix A7, A8 & A9).

39
psuedoF = 4.04;
p = 0.001

0.4

PCoA2

Treatments
10 m
15 m
20 m
Control

0.0

-0.4

-0.3

0.0
0.3
0.6
PCoA1
Figure 2.14 PCoA comparing understory species community composition across all treatment

units two year post timber harvest (July 2020).
psuedoF = 7.52;
p = 0.001

PCoA2

0.4
Treatments
0.0

10 m
Control

-0.4

-0.3

0.0
0.3
0.6
PCoA1
Figure 2.15 PCoA comparing understory species community composition between uncut control

and 10 m thinned treatments two year post timber harvest (July 2020).

40
psuedoF = 5.00;
p = 0.001

PCoA2

0.4
Treatments
15 m
Control

0.0

-0.4

-0.6

-0.3

0.0
PCoA1

0.3

0.6

Figure 2.16 PCoA comparing understory species community composition between uncut control
and 15 m thinned treatments two year post timber harvest (July 2020).
psuedoF = 3.7;
p = 0.001

PCoA2

0.4
Treatments
20 m
Control

0.0

-0.4

-0.4

0.0
PCoA1

0.4

Figure 2.17 PCoA comparing understory species community composition between uncut control
and 20 m thinned treatments two year post timber harvest (July 2020).

41
Strip thinned NDVI and NDRE index products derived from multispectral remote sensing
Additional to ground sampling, forage productivity in the thinned strips of the three adjacent
blocks were monitored using NDVI and NDRE index products post-harvest 2019 and 2020
(Borowik et al. 2013; Kanke et al. 2016) (Figure 2.18). The statistical analysis did not find any
significant different in all thinned strips both post harvest 2019 and 2020 (p>0.05) (Table 2.8).
Though, NDRE analysis conducted in July 2019 did find a difference with a high index product in
15 m and low in 10 m thinned treatments (p<0.05) (Table 2.8).

Figure 2.18 Orthomosaic images with corresponding NDVI and NDRE indexes of the three
adjacent forest sites located in Goudie, East Kelowna, British Columbia. Orthoimages and
multispectral sensor images were taken post timber harvest.

42
Table 2.8: Mean values comparison of understory vegetation greenness and availability
measured by NDVI and NDRE index products derived from a multispectral sensor
between strip-thinned treatments

July 2019
NDVI
NDRE

10 m
0.35 ± 0.014
0.14 ± 0.008

15 m
0.41 ± 0.025
0.18 ± 0.011

20 m
0.4 ± 0.022
0.16 ± 0.012

p-Value
0.5
0.01

August 2019
NDVI
NDRE

10 m
0.53 ± 0.018
0.24 ± 0.006

15 m
0.56 ± 0.022
0.28 ± 0.016

20 m
0.54 ± 0.025
0.26 ± 0.0167

p-Value
0.6
0.15

July 2020
NDVI
NDRE

10 m
0.74 ± 0.015
0.17 ± 0.006

15 m
0.72 ± 0.023
0.17 ± 0.007

20 m
0.73 ± 0.021
0.18 ± 0.009

p-Value
0.8
0.6

Values are means ± standard error (n= 30 for each group). NDVI= Normalized Difference Vegetation index.
NDRE= Normalized Difference Red Edge Index. Treatments means were compared using ANOVA test
(significant level p= 0.05).

43
DISCUSSION
Forage yield and quality
Variation in forage productivity is often observed with the level of thinning. For example, the
productivity is high after timber harvesting and drops dramatically over time depending on the
level of thinning and canopy closure (Thomas et al. 1999; Krzic et al. 2004). However, forage
productivity does not increase immediately due to thinning. Significant growth and development
of forage in open thinned areas can be observed within 10 years (Alaback and Herman 1988).
Forage quantity assessment conducted two years post timber harvesting in our study area found a
high variation in yield between the wider (15 m and 20 m strip widths) and the thinner treatments
(10 m width) as well as the uncut control. A higher biomass was found in the wider areas (15 m
and 20 m strips) and lower biomass found in the thinner areas (10 m strips) as well as in un-thinned
areas (uncut control). This was most likely due to light limitations restricting growth because of
high tree density, the degree of open canopy and the overall basal area along with perhaps soil
moisture in the first meters of the standing forest which all influence understory species
development (McConnell and Smith 1965; Peitz et al. 2001). Levels of temperature and sunlight
absorbed by forage in wider thinned areas stimulate a rapid synthesis of plant cells and accelerate
plant development (Lindgren and Sullivan 2014). However, the effect of thinning across the strip
widths, including un-thinned controls on harvested biomass quality was not significant both postharvest 2019 and 2020. It is possible that the reason for non-differences in CP, soluble
carbohydrate, fat, lignin, TDN, ADF and NDF in all treatment units could be due to the fact that
seeded and native palatable species are not fully established and the fact that unpalatable species
were excluded in our analysis process.
Ishii, Maleque, & Taniguchi, (2008) showed that strip thinned stands have greater influence on
both forage and timber production such as increasing tree DBH (Diameter at Breast Height) than
un-thinned stands. Thinned stands increase the amount of incident light that reaches the forest floor
and reduces nutrients and moisture competition resulting in higher understory forage growth and
development as well as increasing tree stem growth, crown size and timber value, which might
explain some of the outcome obtained.

44
Understory species richness and diversity
Timber harvesting through strip thinning influenced understory species richness and diversity.
As expected, changes in species richness and diversity were observed after two years post-harvest
in July 2020 with high species richness and diversity in 15 m and 20 m thinned treatments. A study
of biodiversity response to intensive biomass production from forest thinning in North American
forests conducted by Verschuyl et al., (2011) found that understory species richness and diversity
frequently respond positively to forest thinning. When we compared the thinned strips and the
uncut control areas one year post timber harvest, the assessment showed a similar plant species
richness and diversity; and greater richness and diversity in the thinned stands than uncut control
stands two years post-harvest, which is consistent with other studies (Ares et al. 2010; Verschuyl
et al. 2011). However, it is possible that both overstory and understory species community
composition pre-harvest 2018 may have influenced the outcome observed post-timber harvest,
especially in the 15 m treatments, as the baseline survey conducted pre-harvest 2018 found a
significant difference between the treatment units with a high species richness and diversity in the
areas assigned to be harvest at 15 m width and low in the areas allocated to 10 m width and uncut
control treatment. A similar trend was also seen post-harvest 2019 and 2020. The analysis of
covariance which assessed the interaction effect between pre-harvest 2018 and post-harvest 20192020 found that both 20 m and 15 m produced high species richness and diversity than uncut
control and 10 m thinned treatments.
Understory species community composition
In addition to the difference in species richness and diversity pre-harvest 2018, species
community composition was different as well among treatment units, both before and after timber
harvesting. Variations in species community composition observed pre-harvest 2018 between the
areas allocated to 15 m strip widths and uncut control treatments were probably due to the
difference in forest density, canopy openings or crown size (Lochhead and Comeau 2012).
According to Halpern & Spies (1995), plant species richness and diversity has long been used
to describe plant species community composition; and thinning significantly influences understory
vegetation community both in the short and long term. A two-year assessment of the effect of strip
thinning on understory plant community showed a significant variation across all treatment units.
Late establishment of plant species across the thinned strips and delay in germination of some
agronomic seed mix added in October 2018 might have played a role in the variation seen both

45
post-harvest 2019 and 2020 assessment. Several mechanisms and factors can influence understory
plant community response post timber harvesting including survival and dispersal rate, succession
and competition for a long-term effects (Gilliam and Roberts 2003).
NDVI and NDRE vegetation index products
Previous studies have shown that NDVI and NDRE index products are commonly used to
predict vegetation growth and availability (Liu et al. 2019; Zhang et al. 2019). In our study, Both
index products were generated from data collected by an Altum multispectral camera, which has
shown to produce highly accurate direct geo-referencing data for plant phenotyping analysis
(Hutton et al. 2020). The results obtained post-harvest 2019 and 2020 did not find any significant
difference between thinned strips at any width. However, the NDRE index product of July 2019
detected a significant change with a higher vegetation growth in the 15 m strip widths and lowest
growth in the 10 m widths. The presence of bare ground and the height difference of understory
vegetation in the open strips might have significantly affected our result as NDVI and NDRE are
both sensitive to soil conditions such as soil moisture, soil structure and topographic features
(Camps et al. 2016; Chen et al. 2019; Ihuoma and Madramootoo 2019).
H. Q. Liu & Huete (1995) mentioned that signal contribution from non-vegetation components
can reduce the accuracy in determining the relationships between individual reflectance bands and
plant parameters, especially during early plant development stages as the stability of these indices
depends on vegetation maturity and coverage.

CONCLUSION
Biodiversity conservation in BC’s managed forests has emphasized how species diversity and
ecological functions can be maintained, while simultaneously improving timber productivity (He
and Barclay 2000). Forest management through an integration of forage, livestock and timber
management approach appears both viable and valuable to promote biodiversity and enhance
overstory and understory vegetation (Udawatta et al. 2019). Our research suggests that the above
integrated management approach can be successfully accomplished through a silvopasture system.
Our study tested the influence of strip thinning at 10 m, 15 m and 20 m widths on forage
productivity. Our hypothesis predicted that the widest strip, thinned at 20 m widths, would
maximize forage productivity better than thinner strips such as the 15 m and 10 m widths used in

46
this study, as well as the un-thinned control areas. In order to determine forage productivity, our
investigation focused on assessing differences in forage growth and development in terms of yield
and quality, species richness, diversity and community composition produced by each strip width
as well as un-thinned control areas. Although the parameters used to determine nutritional value
and digestibility of forage for livestock feed such as CP, soluble carbohydrate, fat, lignin, TDN,
ADF and NDF across all designed treatment units was not influenced by thinning, the yield
obtained provided evidence on the benefit of using strip thinned methods in enhancing forage for
livestock and wildlife. Based on two years assessment, strip thinning contributed to understory
vegetation growth and diversity (Ares et al. 2010).
Forage productivity results obtained post-harvest in 2019 and 2020 partially supported our
hypothesis. In addition to high biomass yields obtained in the widest strips, our results showed that
thinning in general adds to the abundance, richness and diversity of forage. However, our
hypothesis was not fully supported because the 15 m strip widths was seen to produce more species
richness and diversity than the 20 m strip widths. Despite a high species richness and diversity
observed in the 20 m and 15 m thinned strips, continued study is important in order to explore in
depth how thinning affects biodiversity; and how disturbance might influence species diversity
overall in open strips (Verschuyl et al. 2011). And while not significantly different in this particular
study, the use of aerial remote sensing by UAVs showed promise in measuring forage productivity;
and should continue to be explored as a research tool in the future for these types of silvopasture
applications in which data collection using traditional methods is logistically difficult.

47
LITERATURE CITED
Alaback, P.B., Herman, F.R., 1988. Long-term response of understory vegetation to stand
density in Picea-tsuga forests. Can. J. For. Res 1522–1530.
Alfaro, R.I., Akker, L. Van, Hawkes, B., 2015. Characteristics of forest legacies following two
mountain pine beetle outbreaks in British Columbia, Canada. Can. J. For. Res. 45, 1387–
1396. doi:10.1139/cjfr-2015-0042
Anderson, M.J., 2017. Permutational Multivariate Analysis of Variance (PERMANOVA). Wiley
StatsRef Stat. Ref. Online 1–15. doi:10.1002/9781118445112.stat07841
Ares, A., Neill, A.R., Puettmann, K.J., 2010. Understory abundance, species diversity and
functional attribute response to thinning in coniferous stands. For. Ecol. Manage. 260,
1104–1113. doi:10.1016/j.foreco.2010.06.023
Asekova, S., Han, S.I., Choi, H.J., Park, S.J., Shin, D.H., Kwon, C.H., Shannon, J.G., Lee, J.D.,
2016. Determination of forage quality by near-infrared reflectance spectroscopy in soybean.
Turkish J. Agric. For. 40, 45–52. doi:10.3906/TAR-1407-33
Axelson, J.N., Alfaro, R.I., Hawkes, B.C., 2009. Influence of fire and mountain pine beetle on
the dynamics of lodgepole pine stands in British Columbia, Canada. For. Ecol. Manage.
257, 1874–1882. doi:10.1016/j.foreco.2009.01.047
Barbasiewicz, A., Widerski, T., Daliga, K., 2018. The analysis of the accuracy of spatial models
using photogrammetric software: Agisoft Photoscan and Pix4D. E3S Web Conf. 26, 1–5.
doi:10.1051/e3sconf/20182600012
Bonham, C.D., Mergen, D.E., Montoya, S., 2004. Plant Cover Estimation: A Contiguous
Daubenmire Frame. Rangelands 26, 17–22. doi:10.2111/1551501x(2004)26[17:pceacd]2.0.co;2
Borowik, T., Pettorelli, N., Sönnichsen, L., Jędrzejewska, B., 2013. Normalized difference
vegetation index (NDVI) as a predictor of forage availability for ungulates in forest and
field habitats. Eur. J. Wildl. Res. 59, 675–682. doi:10.1007/s10344-013-0720-0

48
Camps, A., Park, H., Pablos, M., Foti, G., Gommenginger, C.P., Liu, P.W., Judge, J., 2016.
Sensitivity of GNSS-R Spaceborne Observations to Soil Moisture and Vegetation. IEEE J.
Sel. Top. Appl. Earth Obs. Remote Sens. 9, 4730–4742.
doi:10.1109/JSTARS.2016.2588467
Canada Government, 2019. Kelowna UBCO, British Columbia. Station operator: ECCC-MSC
[WWW Document]. URL
https://climate.weather.gc.ca/climate_data/daily_data_e.html?hlyRange=2012-1213%7C2020-11-27&dlyRange=2013-12-16%7C2020-1127&mlyRange=%7C&StationID=51117&Prov=BC&urlExtension=_e.html&searchType=st
nProx&optLimit=specDate&Month=5&Day=7&StartYear=1840&EndY
Caradus, J., Woodfield, D.R., Stewart, A. V, 1996. Overview and vision for white clover. Agron.
Soc. New Zeal. Spec. Publ. No. 11. Grassl. Res. Pract. Ser. No. 6 1–6.
Chen, X., Guo, Z., Chen, J., Yang, W., Yao, Y., Zhang, C., Cui, X., Cao, X., 2019. Replacing the
red band with the red-SWIR band (0.74ρred+0.26ρswir) can reduce the sensitivity of
vegetation indices to soil background. Remote Sens. 11, 1–15. doi:10.3390/RS11070851
Conover, W.J., Johnson, M.E., Johnson, M.M., 1981. A comparative Study of Tests for
Homogeneity of Variances, with Applications to the Outer Continental Shelf Bidding Data.
Technometrics 23, 315–361.
Daubenmire, R., 1984. A canopy-coverage method of vegetational analysis. North west Sci. 1–
22.
Deng, L., Mao, Z., Li, X., Hu, Z., Duan, F., Yan, Y., 2018. UAV-based multispectral remote
sensing for precision agriculture: A comparison between different cameras. ISPRS J.
Photogramm. Remote Sens. 146, 124–136. doi:10.1016/j.isprsjprs.2018.09.008
FOSSAnalytical, 2010. InfraXact TM Lab/Pro. WO 2004/027404.
Gilliam, F.S., Roberts, M.R., 2003. response of the herbaceous layer to disturbance in eastern
forests. forests of eastern north america. Oxford University Press. Inc., 198 Madison
Avenue, New york, New york 10016.

49
Goetz, S., Dubayah, R., 2011. Advances in remote sensing technology and implications for
measuring and monitoring forest carbon stocks and change. Carbon Manag. 2, 231–244.
doi:10.4155/cmt.11.18
Halpern, C.B., Spies, T.A., 1995. Plant Species Diversity in Natural and Managed Forests of the
Pacific Northwest. Ecol. Soc. Am. 5, 913–934.
He, F., Barclay, H.J., 2000. Long-term response of understory plant species to thinning and
fertilization in a Douglas-fir plantation on southern Vancouver Island, British Columbia.
Can. J. For. Res. 30, 566–572. doi:10.1139/x99-237
Hutton, J.J., Lipa, G., Baustian, D., Sulik, J., Bruce, R.W., 2020. High accuracy direct
georeferencing of the altum multi-spectral uav camera and its application to high throughput
plant phenotyping. Int. Arch. Photogramm. Remote Sens. Spat. Inf. Sci. - ISPRS Arch. 43,
451–456. doi:10.5194/isprs-archives-XLIII-B1-2020-451-2020
Ihuoma, S.O., Madramootoo, C.A., 2019. Sensitivity of spectral vegetation indices for
monitoring water stress in tomato plants. Comput. Electron. Agric. 163, 104860.
doi:10.1016/j.compag.2019.104860
Insua, J.R., Utsumi, S.A., Basso, B., 2019. Estimation of spatial and temporal variability of
pasture growth and digestibility in grazing rotations coupling unmanned aerial vehicle
(UAV) with crop simulation models. PLoS One 14, 1–21.
doi:10.1371/journal.pone.0212773
Ishii, H.T., Maleque, M.A., Taniguchi, S., 2008. Line thinning promotes stand growth and
understory diversity in Japanese cedar (Cryptomeria japonica D. Don) plantations. J. For.
Res. 13, 73–78. doi:10.1007/s10310-007-0051-7
Jensen, W.A., 2009. Approximations of tolerance intervals for normally distributed data. Qual.
Reliab. Eng. Int. 25, 571–580. doi:10.1002/qre.989
Joshi, C., Leeuw, J. De, Duren, I.C. Van, 2004. Remote Sensing and Gis Applications for
Mapping Spatial Modelling of Invasive Spesies. Geoinf. Sci. 2, 669–677.

50
Kanke, Y., Tubaña, B., Dalen, M., Harrell, D., 2016. Evaluation of red and red-edge reflectancebased vegetation indices for rice biomass and grain yield prediction models in paddy fields.
Precis. Agric. 17, 507–530. doi:10.1007/s11119-016-9433-1
Krzic, M., Page, H., Newman, R.F., Broersma, K., 2004. Aspen regeneration, forage production,
and soil compaction on harvested and grazed boreal aspen stands. BC J. Ecosyst. Manag. 5,
30–38.
Lindgren, P.M.F., Sullivan, T.P., 2014. Response of forage yield and quality to thinning and
fertilization of young forests : implications for silvopasture management 289, 281–289.
Liu, H., Dahlgren, R.A., Larsen, R.E., Devine, S.M., Roche, L.M., O’ Geen, A.T., Wong, A.J.Y.,
Covello, S., Jin, Y., 2019. Estimating rangeland forage production using remote sensing
data from a Small Unmanned Aerial System (sUAS) and planetscope satellite. Remote
Sens. 11. doi:10.3390/rs11050595
Liu, H.Q., Huete, A., 1995. Feedback based modification of the NDVI to minimize canopy
background and atmospheric noise. IEEE Trans. Geosci. Remote Sens. 33, 457–465.
doi:10.1109/36.377946
Lochhead, K.D., Comeau, P.G., 2012. Relationships between forest structure, understorey light
and regeneration in complex Douglas-fir dominated stands in south-eastern British
Columbia. For. Ecol. Manage. 284, 12–22. doi:10.1016/j.foreco.2012.07.029
McConnell, B.R., Smith, J.G., 1970. Response of Understory Vegetation to Ponderosa Pine
Thinning in Eastern Washington. J. Range Manag. 23, 208. doi:10.2307/3896390
McConnell, B.R., Smith, J.G., 1965. Understory Response Three Years after Thinning Pine. J.
Range Manag. 18, 129. doi:10.2307/3895622
Michez, A., Lejeune, P., Bauwens, S., Lalaina Herinaina, A.A., Blaise, Y., Muñoz, E.C., Lebeau,
F., Bindelle, J., 2019. Mapping and monitoring of biomass and grazing in pasture with an
unmanned aerial system. Remote Sens. 11, 1–14. doi:10.3390/rs11050473
Mohd Razali, N., Bee Wah, Y., 2011. Power comparisons of Shapiro-Wilk, Kolmogorov-

51
Smirnov, Lilliefors and Anderson-Darling tests. J. Stat. Model. Anal. 2, 21–33.
Orefice, J., Smith, R.G., Carroll, J., Asbjornsen, H., Howard, T., 2019. Forage productivity and
profitability in newly-established open pasture, silvopasture, and thinned forest production
systems. Agrofor. Syst. 93. doi:10.1007/s10457-016-0052-7
Paliy, O., Shankar, V., 2016. Application of multivariate statistical techniques in microbial
ecology. Mol. Ecol. 25, 1032–1057. doi:10.1111/mec.13536
Pang, X., Bao, W., Zhu, B., Cheng, W., 2013. Responses of soil respiration and its temperature
sensitivity to thinning in a pine plantation. Agric. For. Meteorol. 171–172, 57–64.
doi:10.1016/j.agrformet.2012.12.001
Pavlů, V., Hejcman, M., Pavlů, L., Gaisler, J., Nežerková, P., 2006. Effect of continuous grazing
on forage quality, quantity and animal performance. Agric. Ecosyst. Environ. 113, 349–355.
doi:10.1016/j.agee.2005.10.010
Peitz, D.G., Shelton, M.G., Tappe, P.A., 2001. Forage production after thinning a natural loblolly
pine-hardwood stand to different basal areas. Wildl. Soc. Bull. 29, 697–705.
Pietikäinen, J., Pettersson, M., Bååth, E., 2005. Comparison of temperature effects on soil
respiration and bacterial and fungal growth rates. FEMS Microbiol. Ecol. 52, 49–58.
doi:10.1016/j.femsec.2004.10.002
Schlaepfer, M.A., Sax, D.F., Olden, J.D., 2011. The Potential Conservation Value of Non-Native
Species. Conserv. Biol. 25, 428–437. doi:10.1111/j.1523-1739.2010.01646.x
Schweitzer, S.H., Bryant, F.C., Wester, D.B., 1993. Potential forage species for deer in the
southern mixed prairie. Allen Press behalf Soc. Range Manag. 46, 70–75.
Sullivan, T.P., Sullivan, D.S., Lindgren, P.M.F., 2001. Stand structure and small mammals in
young lodgepole pine forest: 10-year results after thinning. Ecol. Appl. 11, 1151–1173.
doi:10.1890/1051-0761(2001)011[1151:SSASMI]2.0.CO;2
Svejcar, T., Vavra, M., 1985. The Influence of Several Range Improvements on Estimated

52
Carrying Capacity and Potential Beef Production. Soc. Range Manag. 38, 395–399.
Thomas, S.C., Halpern, C.B., Falk, D.A., Liguori, D.A., Austin, K.A., 1999. Plant diversity in
managed forests: Understory responses to thinning and fertilization. Ecol. Appl. 9, 864–879.
doi:10.1890/1051-0761(1999)009[0864:PDIMFU]2.0.CO;2
Tueller, P.T., 1989. Remote sensing applications technology for rangeland management. J.
Range Manag. 42, 442–451.
Udawatta, R.P., Rankoth, L.M., Jose, S., 2019. Agroforestry and biodiversity. Sustain. 11.
doi:10.3390/su11102879
Uniyal, S.K., Awasthi, A., Rawat, G.S., 2005. Biomass availability and forage quality of Eurotia
ceratoides Mey in the rangelands of Changthang, eastern Ladakh. Curr. Sci. Assoc. 89,
201–205.
USDA United States Department of Agriculture, 2019. NRCS- Natural Resources Conservation
Service, Plants Database [WWW Document]. URL https://plants.sc.egov.usda.gov/java/
Verkaik, I., Espelta, J.M., 2006. Post-fire regeneration thinning, cone production, serotiny and
regeneration age in Pinus halepensis. For. Ecol. Manage. 231, 155–163.
doi:10.1016/j.foreco.2006.05.041
Verschuyl, J., Riffell, S., Miller, D., Wigley, T.B., 2011. Biodiversity response to intensive
biomass production from forest thinning in North American forests - A meta-analysis. For.
Ecol. Manage. 261, 221–232. doi:10.1016/j.foreco.2010.10.010
Weidenhamer, J.D., Callaway, R.M., 2010. Direct and Indirect Effects of Invasive Plants on Soil
Chemistry and Ecosystem Function. J. Chem. Ecol. 36, 59–69.
doi:https://doi.org/10.1007/s10886-009-9735-0
Wills, B.J., Douglas, G.B., Mckenzie, J., Trainor, K.D., Foote, A.G., 1998. Thinopyrum
intermedium (Host) Barkw. &amp; Dewey – a review, and evaluation of intermediate and
pubescent wheatgrass for dryland agriculture in New Zealand. New Zeal. Grassl. Assoc. 60,
233–241.

53
Woods, A.J., 2003. Species diversity and forest health in northwest British Columbia. For.
Chron. 79, 892–897. doi:10.5558/tfc79892-5
Woods, A.J., Nussbaum, A., Golding, B., 2000. Predicted impacts of hard pine stem rusts on
lodgepole pine dominated stands in central British Columbia. Can. J. For. Res. 30, 476–481.
doi:10.1139/x99-236
Zaborske, R.R., Hauver, R.N., McClellan, M.H., Hanley, T.A., 2002. Beyond 2001: A
Silvicultural Odyssey to Sustaining Terrestrial and Aquatic Ecosystems- Understory
Vegetation Development Following Commercial Thinning in Southeast Alaska: Preliminary
Results from the Second-Growth Management Area Demonstration Project.
Zeng, L., Chen, C., 2018. Using remote sensing to estimate forage biomass and nutrient contents
at different growth stages. Biomass and Bioenergy 115, 74–81.
doi:10.1016/j.biombioe.2018.04.016
Zhang, K., Ge, X., Shen, P., Li, W., Liu, X., Cao, Q., Zhu, Y., Cao, W., Tian, Y., 2019.
Predicting rice grain yield based on dynamic changes in vegetation indexes during early to
mid-growth stages. Remote Sens. 11. doi:10.3390/rs11040387

54

CHAPTER 3 – EFFECT OF SILVOPASTURE SYSTEMS ESTABLISHED
THROUGH FOREST THINNING ON SOIL CARBON AND NITROGEN
STOCKS.
INTRODUCTION
Soil plays a pivotal role in delivering a wide range of ecosystem services including the support
of agriculture and silvopastoral production (Lavelle et al. 2006). Maintaining soil properties is
critical for achieving additional ecosystem social, economic and environmental needs such as:
provision of food through crop production, timber production, soil water availability, habitat for
living organisms, as well as carbon and nitrogen storage (Smith et al. 2015). Studying soil carbon
and nitrogen storage including its interaction with the earth’s climate system is vital for soil
ecosystem function as nitrogen availability often impacts carbon accumulation especially in
temperate and boreal forests (Sokolov et al. 2008). Recent study of forest ecosystems and the rising
of atmospheric carbon dioxide levels reveals that nitrogen constraints can impact the amount of
carbon sequestered by plant woody materials and decomposition; which can affect the terrestrial
ecosystems and impact the incremental levels of CO2 in the atmosphere (Luo et al. 2004). In
addition, nitrogen content in the soil can be a major factor in stimulating plant growth,
photosynthesis activity, protein production, and the uptake of other nutrients (Novoa and Loomis
1981). Soil carbon is also important and should be well monitored to keep sustaining
agriculture/silvopasture production systems, maintaining and even building resilience for climate
change adaptation while improving soil quality (Lal et al. 2004). Monitoring carbon storage in the
soils is vital as it depends on several factors including: agro-ecological conditions, plant
characteristics, soil characteristics and management practices (Howlett, Moreno, et al. 2011).
Previous studies have highlighted the key benefits of monitoring soil carbon for sustainable forest
management (Fischer et al. 2017); and its advantage in maintaining and contributing to multiple
resources such as: biodiversity, ecosystem services, soil water resources, global ecological cycles
and other social benefits across the same land base (Kneeshaw et al. 2000).
Forest thinning is one of the sustainable forest management strategies capable of managing
multiple concurrent resources (Thomas et al. 1999). Forest thinning increases the resistance and
resilience from natural disturbances such as, wildfires and forest insect pests (Hood et al. 2016).
Such pests like mountain pine beetle (Dendroctonus ponderosae) has impacted up to 10.1 million

55
hectares of British Columbia’s lodgepole pine (Pinus contorta) forest (Axelson et al. 2009) and
resulted in changing forest carbon dynamics and increases in atmospheric carbon emission. The
pine beetles has been estimated to impact an estimated amount of approximately 270 megatonnes
(Mt) of forest carbon net release over the last two decades, or 36 g carbon m-2 yr-1 on over 374,000
km2 of BC forested land (Kurz et al. 2008).
Integrating silvopasture system into forest management after thinning can be an additional tool,
offering extra benefits in maximizing land use, including forage and timber productivity as well
as carbon sequestration above and belowground level (Howlett, Mosquera-Losada, et al. 2011).
The thinning method contributes to amending soil temperature, reducing competition, improving
light penetration and soil water content (Saunders et al. 2012). Silvopasture, an integrated forage,
livestock and timber management strategy (Shrestha and Alavalapati 2004) which introduces
additional grass biomass due to canopy opening; is considered one of the most efficient methods
to sequester and store carbon above and belowground (Scurlock and Hall 1998; Feliciano et al.
2018). Grasses store a significant portion of carbon belowground while forested ecosystems tend
to accumulate a large portion of carbon aboveground (Post and Kwon 2000).
Despite the value of silvopasture practice for soil carbon sequestration, and as an approach to
optimize multiple ecosystem services across the landscape, more research beyond forage and
timber production is required on silvopastoral systems on the interaction between livestock grazing
and soil (Sharrow 2007). Livestock grazing is a common practice which has been traditionally
used, and currently implemented to support people’s livelihoods (Gurung et al. 2009). However,
overgrazing can create soil exposure to several disturbances, deteriorate soil organic matter (SOM)
(Conant and Paustian 2002) and dramatically alter the potential for silvopasture systems to
sequester carbon (Silveira et al. 2014; Richardson et al. 2017).
Previous research by Pineiro et al. (2010) tested the effect of disturbance, including grazing, on
Soil Organic Matter (SOM) stocks considered as a major reservoir of soil organic carbon (SOC);
and soil organic nitrogen (SON) and its impact on soil water availability as well as soil fertility
and structure. Referring to their figure (3.1) below, their findings revealed that overgrazing
modifies the process of SOM decomposition by changing the net primary production ratio that
reaches the soil, which results in reducing SOC and SON stocks.

56

Figure 3.1: “Soil organic carbon (SOC); a) and soil organic nitrogen (SON); b) controls at
different temporal scales. Dashed lines show which controls are affected by grazing. ANPP is
aboveground net primary production, and BNPP is belowground net primary production”
(Pineiro et al. 2010).

57
However, effective grazing management strategies, for example light to moderate grazing, can be
a key in improving the accumulation of soil carbon storage by retaining a mix of forage species
community composition, increased root production and facilitating soil development (Frank et al.
1995; Richardson et al. 2017).
In our study, we focused on assessing the effects of the silvopasture model of forest
management through strip thinning by deploying three different strip widths (10 m, 15 m and 20
m thinned widths), along with the uncut control, on soil organic matter (SOM), soil total carbon
(TC) and total nitrogen (TN) as well as the change in soil pH, bulk density and CO2 flux through
soil respiration due to canopy opening. Data was collected both before and after strip thinning of
a lodgepole pine forest in three adjacent forested areas; located in the southern interior of BC in
order to test the hypothesis that strip thinning would have a significant impact on overall soil
carbon and nitrogen stocks.

MATERIALS & METHODS
Site Description & Research Design
Referring to the study site in the Materials & Methods of Chapter 2, the field experiments were
conducted in the same area throughout three consecutive seasons (pre-timber harvest 2018, postharvest 2019 and 2020) in the three adjacent forested areas of Goudie, Kelowna, British Columbia
situated at an elevation range between 1340 – 1400 m (Fig 2.1). For data collection purposes, two
strips were selected in 10 m, 15 m, and 20 m strips, including two plots in the uncut control areas.
The sample data were collected from the center areas of each of the selected strips toward north
for strips facing north and east for strips facing east.
Soil sampling for bulk density analysis
Post-harvest 2019 sampling locations were established in each treatment unit for assessing soil
compaction caused by heavy machinery during strip thinning activity. A total of 24 samples were
collected in the middle areas of the treatment (2 samples per treatment unit) using a core method
(Rab 1994). The treatments were re-assessed a year later in June 2020 targeting the entire thinned
strip. A two dimensional transect system was utilized to establish random sampling locations
throughout each treatment unit. A 90 m transect was placed parallel at 1 m distance from the edge
of the strip followed by a perpendicular transect placed across the strip width. 16 sampling

58
locations were randomly selected in each treatment unit (8 sampling points per strip) using both
transects which were established lengthwise and widthwise in both open strips and uncut control.
At each location, litter, decayed material, and mineral soil samples were collected separately using
an excavation method. Mineral soil for bulk density measurement was collected at each location
to a soil depth of 15 cm including rocks and other materials found in the excavated plot. A total of
192 soil samples were collected across the three blocks, weighed and placed in a drying oven at
70oC for 96 hours. Dried samples were weighed again and then sieved to remove rocks and other
materials larger than 2 mm in diameter. Sieved samples were weighed and the bulk density was
calculated based on the mass-volume ratio (Blaisdell et al. 2003; Maynard and Curran 2006).
Soil sampling for pH, SOM, TN & TC analysis
Prior to timber harvesting, soil samples were collected in July 2018 from the four treatments
(10 m, 15 m, 20 m strip widths and uncut control) in order to determine soil properties baseline
information in each of the three forested areas. Two sampling strips per each thinned treatment
unit were selected including two sampling areas in the uncut control (Figure. 2.1). A maximum of
10 sampling points per treatment were recorded and the soils were collected separately from two
different depths (0-10 cm and 10-20 cm soil depth). Five samples above 10 cm soil depth and five
below 10 cm up to 20 cm soil depth were kept and analyzed separately.
In a similar manner to the July 2018 soil sampling approach, soils were collected post-harvest
in July 2019 and July 2020. To better assess the influence of strip thinning at different widths, a
total of 480 soil samples were collected (240 samples collected above 10 cm soil depth and 240
samples between 10 cm and 20 cm depth) in the middle areas of the strips using a 50 m transect.
The edge effect and road effect were excluded in the sampling procedure as both road disturbance
and conifer plantations directly and indirectly influence biological and chemical properties of the
soil including: soil carbon dynamics, pH, organic matter, TC and TN (Hofmeister et al. 2013;
Deljouei et al. 2018).
Soil pH
Pre-harvest 2018 soil collected in July was analyzed to assess soil chemical property baseline
information of the three adjacent blocks. Post-harvest July 2019 and July 2020 soils were collected
in each treatment unit to assess the influence of strip thinning at 10, 15 and 20 m widths on soil

59
acidity or alkalinity. 10 g of fresh soil was placed in 50 ml falcon tubes and mixed with 25 ml of
distilled water. The mixture was used to measure soil pH with a Palintest 800 PT1350 pH meter
after shaking the soil water for 1 min and let it rest for 60 min.
Soil Organic Matter (SOM)
Soil organic matter content was determined using the loss on ignition (LOI) method. 1.5 g of
fresh soils were placed into aluminum tin foil pans and heated at 105oC for 12 hours using a
YAMATO forced convection constant temperature drying oven (DKN818, Yamato Scientific Co.
Ltd) in order to remove soil moisture. The soils were then weighed on an analytical scale and the
weights recorded before placing the dried samples into the muffle furnace. The Barnsteadthermolyne 62700 furnace was used to ignite the soils at 500oC for 5 hours and then left in the
desiccator for at least 30min until reached the room temperature. Finally, the samples were
weighed and recorded again. The soil organic matter was calculated using (Wang et al. 2011; Wang
et al. 2012):
Eq. (2)
SOMLOI (gkg-1) =

𝑊𝑒𝑖𝑔ℎ𝑡105𝐶 − 𝑊𝑒𝑖𝑔ℎ𝑡 500𝐶 ×1000
𝑊𝑒𝑖𝑔ℎ𝑡105𝐶

SOM final results in gkg-1 were finally converted to % for comparison with the total carbon results.
Soil total nitrogen (TN) and total carbon (TC)
Soil samples were carefully prepared using aluminum tin foil pan. Samples were ground and
sifted through a 355µm laboratory test sieve (mesh No 45). The sieved samples were weighed in
tin containers and then introduced into the combustion reactor from the FlashSmart Elemental
Analyzer (Thermo Fisher scientific TM) (Figure 3.2). The FlashSmart Analyzer function based on
the dynamic flash combustion technique was used, the produced gases after combustion were
carried out by helium flow to the copper reactor, then through a water trap, and finally detected by
the Thermal Conductivity Detector (TCD) (Krotz et al. 2016).

60

Figure 3.2: Thermo Fisher FlashSmart Elemental Analyzer Nitrogen and Carbon configuration
and analysis process (Krotz et al. 2016).
Net Carbon Exchange (NCE) through soil respiration
At the ecosystem level, It is important to understand the Net Carbon Exchange (NCE) between
the atmosphere and the ecosystem including above and below ground activities such as plant
photosynthesis, plant and soil respiration and decomposition (Bhattarai et al. 2016). CO2 is
sequestered through photosynthetic carbon metabolism and utilized for growth and development
of plants. Throughout the fixation process, some carbon content is lost through respiration (LiCOR Biosciences 2010a). Several methods can be used to assess the impact of thinning on the rate
of CO2 release from the ground to the atmosphere and our experiment used an automated LI-COR
8100 Soil CO2 flux system (Figure 3.3) as an approach to quantify CO2 flux through soil or plant
respiration.
The LI-8100 is the most common direct-survey measurement method, using an open and closed
chamber system that analyses the fluxes of CO2 (Fc, μmol m-2s-1) from the soil atmosphere out
into the bulk atmosphere using the equation (3) (Madsen et al. 2009). The system requires the use
of a soil collar (made in polyvinyl chloride- PVC pipe) that is permanently inserted in the soil at
least 24 hours prior to conduct any measurement (Figure 3.3). Two strips per each thinned
treatment unit including uncut control areas were selected for seasonal flux measurement (see
selected strips on figure 2.1). The total of 24 surveys were conducted three times in spring, summer
and fall of 2019 and 2020. For the purpose of insuring the accuracy of the instrument data
collection, repeated measurements at each sampling location were conducted three times and the
average value was recorded using SoilFluxPro software (Li-COR Biosciences 2010b).

61

Eq. (3) (Madsen et al. 2009)

Fc = Fluxes of CO2 (in μmol m-2s-1)
P = Atmospheric pressure (Pa),
V = Total system volume (in m3), including the volume of the chamber, the pump, and tubing in
the measurement loop,
R = Gas constant (8.314 Pa m3 °K-1 mol-1),
T = Absolute temperature (°K), and
S = soil area covered by the chamber (m2).

Figure 3.3: LI-COR, LI-8100A automated CO2 soil gas flux system with a direct survey chamber
ranged between 0ppm to 20,000ppm and an operating temperature range between -20oC to 45oC.
A green soil Collar with a 10 cm inside and 11.4 cm outside diameter, and with 10 cm of length
inserted permanently in the soil, and a minimum of 2 cm extension above the soil surface (Li-COR
Biosciences 2010a).

STATISTICAL ANALYSIS
All statistical analyses were conducted using R version 3.6.2 (The R Foundation for Statistical
Computing). All data sets were checked for normality using Shapiro test and residual plots (Mohd
Razali and Bee Wah 2011). Homogeneity of variance was assessed using the Fligner-Killeen test

62
(Conover et al. 1981), and when necessary, the data were transformed using a natural logarithm or
a square root function (sqrt). Analysis of variance was used to assess the differences in SOM, TC,
TN, pH level, bulk density and the rate of CO2 movement between 10 m, 15 m, 20 m strip thinned
and the uncut control treatments. Analysis of variance was followed by a Tukey post-hoc test to
find treatments that were significantly different from each other at a 5% probability level. Kruskal
Wallis tests followed by Dunn’s post-hoc test, using the Benjamini-Hochberg procedure to control
for multiple comparisons were applied to the data that did not follow a normal distribution and
equal variance assumptions. Analysis of covariance (ANCOVA) was also employed to control for a
priori effect pre-timber harvest 2018.

RESULTS
Soil bulk density
Soil bulk density measurement conducted in the middle areas of the thinned strips (10 m, 15 m
and 20 m width), including the uncut control treatment, was done in May 2019 after soil horizon
and texture surveys, as well as coarse fragments content collected in October 2017 (AppendixA.10& A.11). No effect of heavy machinery utilization on soil bulk density was observed in the
middle areas of the thinned strips (p > 0.05) (Figure 3.4). In May 2020, soil bulk density was
surveyed in the entire strips. The assessment found a significant effect of heavy machinery on soil
compaction (p < 0.001), with a high bulk density in all thinned strips (Figure 3.4). Uncut controls
were significantly different with low bulk density than all of the thinned strip treatments (p< 0.05).
20 m and 10 m strips were also significantly different than 15 m strips (p< 0.05). No difference
was found between 10 m and 20 m strips (p> 0.05).

63

Figure 3.4: Post-harvest 2019 and 2020 mean bulk density (BD in g cm-3) of each treatment unit
in three adjacent block design. Error bars represent standard error of the mean (n= 6 for each
group in 2019 and 48 in May 2020). (NS= non-significant; * = p≤ 0.05; ** = p≤ 0.01; *** = p≤
0.001; **** = p≤ 0.0001).
Soil pH
The average soil pH of the three blocks in 2018 was 5.6 from 0-10 cm depth and 5.7 from 1020 cm depth. The mean soil pH for both depths (0-10 cm and 10-20 cm) did not show any
statistically significant difference between the treatment units (p > 0.05) (Figure 3.5 & 3.6).
Apparently, soil pH in the three blocks slightly reduced by 0.1, from year 2018 to year 2019, as
the average soil pH in 2019 was 5.5 from 0-10 cm depth and 5.61 from 10-20 cm depth. Postharvest 2019 thinning activity had a significant effect on soil pH analyzed above 10 cm soil depth
(p < 0.01) with a high pH in 20 m, followed by the 15 m and 10 m thinned strip treatments and
was very low in the uncut control treatment (Figure 3.5); but no significant effect was observed on
soil pH depth from 10-20 cm between the three different thinned strip treatments and the control
treatment (p > 0.05) (Figure 3.6). A similar assessment was conducted again in July 2020 and the
analysis did not find any significant different in both above and below 10 cm soil depth (p> 0.05)

64
(Figure 3.5 & 3.6). ANCOVA assessment which included pre-harvest 2018 results as a covariate
did not find any effect of pre-harvest condition on soil pH post-harvest 2019 and 2020 both 0-10
cm and 10-20 cm soil depth (p> 0.05).

Figure 3.5: Mean soil pH above 10 cm depth of each treatment unit in three adjacent block designs
pre-harvest 2018, post-harvest 2019 and 2020. Error bars represent standard error of the mean
(n=30 for each group). (NS= non-significant; * = p≤ 0.05; ** = p≤ 0.01; *** = p≤ 0.001; ****
= p≤ 0.0001).

65

Figure 3.6: Mean soil pH below 10 cm depth of each treatment unit in three adjacent block designs
pre-harvest 2018, post-harvest 2019 and 2020. Error bars represent standard error of the mean
(n=30 for each group). (NS= non-significant; * = p≤ 0.05; ** = p≤ 0.01; *** = p≤ 0.001; ****
= p≤ 0.0001).
SOM
The mean SOM level pre-harvest 2018 was at 30.2% and 14.03% from 0-10 cm and 10-20 cm
depth respectively in each of the three adjacent blocks. A statistical comparison above 10 cm soil
depth between areas allocated to 10 m, 15 m, and 20 m thinned treatments, along with the uncut
control treatment, showed a significant difference after a Kruskal Wallis test; as the data was not
normally distributed even after performing various transformation tests (p < 0.05) (Figure 3.7).
Dunn’s test by Benjamini-Hochberg observed a high SOM content in areas allocated to the 15 m
thinned strip treatments and a low content in areas allocated to the 10 m, and 20 m thinned and the
uncut control treatments. No significant difference was observed below 10 cm soil depth (p. 0.05)

66
(Figure 3.8); but Dunn’s test by Benjamini-Hochberg observed a slightly higher SOM content
below the 10 cm soil depth in the areas designed for 15 m thinned treatments and low in the 20 m
thinned treatments (p< 0.05).
In 2019 post-harvest, the mean SOM decreased by 23.7% and 10.23%, as the mean values were
6.5% above 10 cm soil depth, and 3.8% below 10 cm soil depth. SOM content above 10 cm soil
depth was significantly different, and Dunn’s test by Benjamini-Hochberg observed a higher SOM
content in the uncut control treatment than 15 m and 20 m thinned strip treatments and slightly
lower in 10 m thinned treatments (p < 0.05) (Figure 3.7). The analysis of SOM content assessed
below the 10 cm soil depth did not find any thinning effects across all the treatment units (p> 0.05)
(Figure 3.8). Similar analysis was conducted once again in July 2020, two years after thinning
activities. The overall mean SOM content above the 10 cm soil depth was 12.46% and 7.57%
between 10-20 cm depth. Apparently, the mean SOM in 2020 increased by 5.96% above 10 cm
soil depth, and 3.73% below the 10 cm soil depth after comparing with the previous year. KruskalWallis did not find any thinning effect on SOM in any thinned treatment at any soil depth including
the uncut control treatment (p> 0.05) (Figure 3.7 & 3.8). ANCOVA assessment which included
pre-harvest 2018 results as a covariate did not find any effect of pre-harvest condition on SOM
post-harvest 2019 and 2020 both 0-10 cm and 10-20 cm soil depth (p> 0.05).

67

Figure 3.7 Mean SOM (Soil Organic Matter) above 10 cm depth of each treatment unit in three
adjacent block designs pre-harvest 2018, post-harvest 2019 and 2020. Error bars represent
standard error of the mean (n=30 for each group). (NS= non-significant; * = p≤ 0.05; ** = p≤
0.01; *** = p≤ 0.001; **** = p≤ 0.0001).

68

Figure 3.8: Mean SOM (Soil Organic Matter) below 10 cm up to 20 cm soil depth of each
treatment unit in three adjacent block designs pre-harvest 2018, post-harvest 2019 and 2020.
Error bars represent standard error of the mean (n=30 for each group). (NS= non-significant; *
= p≤ 0.05; ** = p≤ 0.01; *** = p≤ 0.001; **** = p≤ 0.0001).

Total Carbon (TC)
Pre-harvest 2018 soil TC content above the 10 cm soil depth were 6.31% and 3.35% below 10 cm
up to 20 cm soil depth. Soil carbon above 10 cm soil depth showed a significant difference, with
a higher carbon content in the areas allocated to the 15 m thinned treatments and lower in other
areas allocated to the 20 m and 10 m thinned treatments (p < 0.05) (Figure 3.9). Uncut control
treatment was also significantly different and higher than the areas allocated to the 20 m thinned
treatments in the top soils above 10 cm depth (p< 0.05) (Figure 3.9). Comparison conducted below
10 cm soil depth with Dunn’s test by Benjamini-Hochberg observed again a higher carbon content
in the areas allocated to the 15 m than 20 m and 10 m thinned treatments (p< 0.05) (Figure 3.10).
Post timber harvest 2019 carbon content was 4.93% above 10 cm soil depth and 2.61% between
10 cm and 20 cm soil depth. No significant effects of thinning at different strip width were

69
observed on %TC at any soil depth (0-10 cm and 10-20 cm) after comparing 10 m, 15 m and 20
m thinned strip treatments including the uncut control (p > 0.05) (Figure 3.9 & 3.10). However, a
similar monitoring approach conducted in July 2020 (average TC content of 4.21% above 10 cm
and 2.7% below 10 cm soil depth) found a significant change in soil TC level. The change was
seen in the deeper soils between 10 cm and 20 cm soil depth, with a higher carbon content in 10
m thinned treatments and lower in 20 m thinned treatments (p<0.05) (Figure 3.10). Although the
top soils above 10 cm depth was slightly different between 10 m and 20 m treatment widths, no
statistically significant difference was observed (at a p> 0.05) (Figure 3.9). Analysis of covariance
which included pre-harvest 2018 results as a covariate did not find any effect of pre-harvest
condition on TC post-harvest 2019 and 2020 both above and below 10 cm soil depth (p> 0.05).

Figure 3.9: Mean TC (Total Carbon) above 10 cm soil depth of each treatment unit in three
adjacent block designs pre-harvest 2018, post-harvest 2019 and 2020. Error bars represent
standard error of the mean (n=30 for each group). (NS= non-significant; * = p≤ 0.05; ** = p≤
0.01; *** = p≤ 0.001; **** = p≤ 0.0001).

70

Figure 3.10: Mean TC (Total Carbon) below 10 cm up to 20 cm soil depth of each treatment unit
in three adjacent block designs pre-harvest 2018, post-harvest 2019 and 2020. Error bars
represent standard error of the mean (n=30 for each group). (NS= non-significant; * = p≤ 0.05;
** = p≤ 0.01; *** = p≤ 0.001; **** = p≤ 0.0001).

Total Nitrogen (TN)
Pre-harvest 2018 soil TN content ranged between 0.23% above 10 cm and 0.14% below 10 cm
soil depth. The comparison between the four treatments on a soil depth of 10 cm showed a high
nitrogen content in the treatments assigned to be harvested at 15 m strip widths and low in other
treatments assigned to be harvested at 10 m and 20 m strip widths (p< 0.05) (Figure 3.11). The
uncut control treatment was also significantly different and had higher TN content than the
treatment intended to be the 20 m thinned treatment (p< 0.05) (Figure 3.11). Nitrogen content
below 10 cm soil depth was still higher in the treatment intended to be the 15 m thinned treatment
and lower in 20 m and 10 m thinned treatment (p< 0.05) (Figure 3.12).
A similar monitoring approach was conducted post-harvest 2019, which showed a slight
variation in nitrogen content with 0.21% above 10 cm soil depth and 0.15% between 10 cm and

71
20 cm soil depth; but no significant effect of timber harvesting at any strip width treatment
including the uncut control treatment was observed either above or below 10 cm soil depth (p>
0.05) (Figure 3.11 & 3.12). However, the re-assessment conducted post-harvest 2020 showed an
increment of nitrogen content above and below 10 cm soil depth of 0.81% and 0.17% respectively.
A significant effect of timber harvesting was observed above 10 cm soil depth (p< 0.05) with a
higher nitrogen content in 10 and 15 m thinned treatments and lower in 20 m thinned treatment
and uncut control treatment (Figure 3.11). No effect was noticed in the deeper soil between 10 cm
and 20 cm soil depths (p> 0.05) (Figure 3.12). Analysis of covariance found a significant
difference only in the top soil (0-10 cm soil depth) (p< 0.05), with a higher TN content in 15 m
and 10 m thinned treatments and a lower nitrogen content in 20 m and uncut control treatments
(Table 3.1).

Figure 3.11: Mean TN (Total Nitrogen) above 10 cm soil depth of each treatment unit in three
adjacent block designs pre-harvest 2018, post-harvest 2019 and 2020. Error bars represent
standard error of the mean (n=30 for each group). (NS= non-significant; * = p≤ 0.05; ** = p≤
0.01; *** = p≤ 0.001; **** = p≤ 0.0001).

72

Figure 3.12: Mean TN (Total Nitrogen) below 10 cm up to 20 cm soil depth of each treatment unit
in three adjacent block designs pre-harvest 2018, post-harvest 2019 and 2020. Error bars
represent standard error of the mean (n=30 for each group). (NS= non-significant; * = p≤ 0.05;
** = p≤ 0.01; *** = p≤ 0.001; **** = p≤ 0.0001).
Table 3.1: ANCOVA and post-hoc test results for soil total carbon (0-10 cm soil depth) preharvest 2018 and post-harvest 2019-2020.

15 m - 10 m
20 m - 10 m
Control - 10 m
20 m - 15 m
Control - 15 m
Control - 20 m

% Soil TN
Estimate Std. Error
-0.03262 0.1188
-0.42047 0.1188
-0.39104 0.1188
-0.38784 0.1188
-0.35842 0.1188
0.02943
0.1188

t value
-0.275
-3.539
-3.292
-3.265
-3.017
0.248

Pr(>|t|)
0.99276
0.00267 **
0.00611 **
0.00641 **
0.01464 *
0.99465

Net Carbon Exchange (NCE)
CO2 flux surveyed post-harvest 2019 averaged 2.63 μmol m-2s-1 in May 2019, and 3.64 μmol
m-2s-1 in July 2019, and finally 0.63μmol m-2s-1 in October 2019. There was no impact of thinning

73
observed in any strip thinned treatment on CO2 flux movement in May and July after comparing
the four treatments (p > 0.05) (figure 3.13). However, a significant impact was observed in October
between the uncut control and 10 m thinned treatments with a higher flux in the uncut control and
lower in the 10 m thinned treatments (p < 0.05) (Figure 3.13).
Post-harvest 2020, CO2 flux re-assessed in May averaged 0.39 μmol m-2s-1, July flux averaged
5.28 μmol m-2s-1, and 0.79 μmol m-2s-1 in October 2020. The comparison analysis did not find any
significant difference across treatments in all seasons (spring, summer and fall) (p> 0.05) (Figure
3.14).
fall_2019

spring_2019

summer_2019

ns

ns

ns

10.0

ns

ns
ns

ns
ns

ns

ns
7.5

ns

ns

ns

**
Anova, p = 0.014

CO2 Flux

ns
ns

ns
Anova, p = 0.43

ns
Anova, p = 0.15

5.0

2.5

0.0
10 m

15 m

20 m

Control

10 m

15 m

20 m

Control

10 m

15 m

20 m

Control

Treatments

Figure 3.13: Mean CO2 Flux content of the treatment units in three adjacent study blocks collected
in May 2019, July 2019 and October 2019. Error bars represent standard error of the mean (n=
6 for each group). (NS= non-significant; * = p≤ 0.05; ** = p≤ 0.01; *** = p≤ 0.001; **** = p≤
0.0001).

74

fall2020co2flux

spring2020co2flux

summer2020co2flux

ns

ns

ns

ns

15

ns
ns

ns
ns

ns

ns

ns

ns

ns

*
Anova, p = 0.092

*

ns
Anova, p = 0.95

ns
Anova, p = 0.084

CO2 Flux

10

ns

5

0
10 m

15 m

20 m

Control

10 m

15 m

20 m

Control

10 m

15 m

20 m

Control

Treatments

Figure 3.14: Mean CO2 Flux content of the treatment units in three adjacent study blocks collected
in May 2020, July 2020 and October 2020. Error bars represent standard error of the mean (n=
6 for each group). (NS= non-significant; * = p≤ 0.05; ** = p≤ 0.01; *** = p≤ 0.001; **** = p≤
0.0001).

75
DISCUSSION
Effect of strip thinning on soil physical and chemical properties
Bulk density and pH
Strip thinning activity in all three adjacent blocks was carried out using heavy machinery and
large tractors, which significantly disturbed the soil in the thinned strips. Soil compaction was
visually observed in all harvested strips after an aerial multispectral sensor scanned all
experimental research blocks (Appendix- A.13). Bulk density assessment of soil compaction
conducted in the middle areas of the thinned strips post timber harvest 2019 did not detected any
compaction issue between the treatment units. However, the second assessment that covered the
entire strips post-harvest 2020 found a high compaction in all the open strips but the 20 m strips
were more affected by compaction, suggesting the 20 m open strips were used more as paths for
transporting timber due to their width. A study by Picchio et al. (2012) showed that although
thinning is an important practice in managing forest products, it can create several modifications
to soil properties including physical properties such as soil compaction which can be caused by
the use of heavy machinery and large tractors.
In addition to the above soil physical property, variation in soil chemical properties were noticed
after strip thinning post-harvest 2019 with an increment of soil pH in all thinned treatments after
comparing them with the uncut control treatment. Lodgepole pine forests have been shown to play
a significant role in changing soil chemical properties especially in increasing soil pH levels in an
open canopy area, while keeping a low pH in a dense and closed canopy area (Vesterdal and
Raulund-Rasmussen 1998; Dingaan et al. 2017). Cheng et al. (2013) found that these variations in
soil pH between forested areas and grass areas are caused by the differences in physical and
chemical characteristics of soil; and these differences can be seen in a short period of time after
land-use change. The increase of soil pH has a greater contribution to forage production, by
increasing the availability of nutrients and nodule formation on leguminous species, such as white
clover (Rice et al. 1977), and enhancing plant community composition and species richness
(Vesterdal and Raulund-Rasmussen 1998; Dingaan et al. 2017). Despite a rapid change of soil pH
level due to lodgepole pine (Pinus contorta) canopy openings (Vesterdal and Raulund-Rasmussen
1998), thinning activities in forested areas tend to affect more of the soil surface than deeper soils;
and our results indicated that soil pH variation was noticed more above 10 cm than below 10 cm

76
soil depth; probably due to lesser amounts of decayed organic matter in top soil, and less
disturbance in deeper soils (Zhang et al. 2017). However, soil pH analyzed two years after thinning
showed a balanced pH level in thinned and un-thinned treatment units. It seems that two years
later, strip thinning did not affect soil pH at any strip width, both above and below the 10 cm soil
depths, and it is possible that understory forage growth, and the addition of soil organic matter,
might have influenced the results obtained.
Soil Organic Matter (SOM)
Soil pH is often considered as a major influencer in regulating SOM turnover (Kemmitt et al.
2006). A significant increase of soil pH in open strips, and litter clearing by heavy machinery
during the process of timber harvesting, might have contributed to a significant reduction of SOM
content in 20 m, 15 m and 10 m thinned treatments post-harvest 2019 (Turner and Lambert 2000).
According to Turner & Lambert (2000), thinning increases SOM and carbon content after
vegetation has fully established through root turnover, rapid litter and root decomposition. Soil
disturbance during thinning, and removal of forest floor cover, can negatively affect SOM and
carbon content especially above 10 cm soil depth through a slow litter accumulation which
explains a significant low organic matter content observed in 20 m, 15 m and 10 m thinned
treatments compared to the uncut control treatment. It has been shown that land-use change by
opening forest canopies, prior to understory forage growth, reduces SOM and SOC by 42 to 59%,
and once vegetation is fully established, accumulation of organic carbon increases at a rate of 300350kg C ha-1 year-1 (Puget and Lal 2005). This might explain an increment of SOM and the nonsignificant differences observed across treatment units in July 2020. Although, understory
vegetation in the open strips is not yet fully established, vegetation growth in July 2020 was
significantly higher than what was observed in July 2019. A significant growth and the subsequent
decay process of plant litter in an agroforestry area could ultimately be a main factor in the
formation of soil organic matter (Melillo et al. 1989), and as such vegetation growth needs to be
monitored in the future.
Total Carbon (TC) and Nitrogen (TN)
The dominance of conifer tree species in some areas of a forested stand can affect soil carbon
and nitrogen as observed pre-harvest 2018; with a significant variation across the designed
treatment units (Finzi et al. 1998). It is possible that the difference in litter production, and rate of
litter decomposition, due to the amount of tree canopy cover, soil temperature and moisture may

77
have also affected the result obtained pre-harvest 2018 where some of the areas designed for 15 m
thinned treatments and control treatments had a significantly higher TC and TN content than some
areas designed for 10 m and 20 m thinned treatments (Finzi et al. 1998). However, thinning
activities across treatment units post-harvest 2019 did not show any influence on %TC and %TN
content; most likely due to the fact that some of the thinned litter materials remained in the open
strips, and ended up mixing with the soil after decomposition, like the study published by Bai et
al. (2017). The addition of the seed mix and the plant establishment process in the open areas might
have added more litter and influenced microbial activities.
Results from this study obtained post-harvest 2020 showed that, assessing long-term effects of
forest management on ecosystem pools, including changes in soil carbon and nitrogen, after forest
tree removal might take several years (Grady and Hart 2006). The response of soil physical and
chemical properties post thinning might be seen in various ways depending on the study sites and
assessment periods (Wic Baena et al. 2013). For example, similarities in amounts of soil carbon
were observed between treatment units post-harvest 2019 and 2020; as were differences in soil
nitrogen with an incremental amount in 10 m and 15 m rather than 20 m thinned treatments and
un-thinned control treatments. Soil compaction and high plant root activities, can also affect soil
properties such as aeration, water content, temperature and soil microbial activities; especially in
sandy loam soils which can influence soil carbon and nitrogen mineralization and nitrification of
soil organic matter (Brevik et al. 2002). (Appendix- A.11).
Net Carbon Exchange (NCE)
As mentioned above, the effect of forest thinning on soil respiration can be determined by
different inter-related factors including: changes in climatic conditions, soil temperature, microbial
respiration, root respiration and decomposition rate from dead branches, and litter to root materials
(Tang et al. 2005). In our study, we quantified soil CO2 flux throughout different seasons (spring,
summer and fall seasons) in order to account for the different factors mentioned above. Carbon
flux was determined after placing permanent collars (Fig 3.3) in the center areas of the four
treatments (uncut control, 10 m, 15 m and 20 m thinned treatments), across the total 101.4 hectares
of our three adjacent study blocks. In the three seasons monitored (spring, summer and fall) postharvest between 2019 and 2020, a significant effect of strip thinning on CO2 flux was found only
in the fall of 2019, with a higher flux level in the un-thinned control treatments, and a lower flux
in 10 m thinned treatments. Although, CO2 flux monitored in all three seasons post-harvest 2020

78
did not find any significant difference at p< 0.05, uncut control treatments showed a slight higher
flux than 15 m thinned treatments in summer, and higher than 10 m thinned treatments in fall.
Olajuyigbe et al. (2012) showed that vegetation growth and biomass productivity have a strong
influence on microbial activities and soil respiration in active trees rather than in dead organic
material. Additionally, Tang et al. (2005) showed how delay in vegetation growth is not only
explained by soil temperature and soil water, but also by variations in root biomass, ground
vegetation cover and soil properties. It seems that the uncut control treatments provided a high
flux than thinned areas. It is more likely that the combination of higher forest tree, root biomass
and microbial activity, along with soil temperature and other factors mentioned previously might
have also significantly influenced our results.

CONCLUSION
This study provides deeper insight on whether the adoption of silvopasture systems established
through strip thinning treatments at 10 m, 15 m or 20 m widths respectively, can potentially
improve soil carbon sequestration as well as nitrogen storage, ultimately creating more productive
and resilient ecosystems for livestock and wildlife. Our findings were not entirely consistent with
our hypothesis, and the results obtained did not fully support that the 20 m thinned treatments will
sequester more soil carbon and provide more nitrogen than uncut controls or 10 m and 15 m
thinned treatments. In contrast, the 15 m and 10 m thinned treatments showed a tendency to
accumulate more carbon and nitrogen content than the 20 m thinned treatments. As explained
previously in the discussion section, variation in soil pH, soil disturbance and loss of soil organic
matter post timber harvesting had a positive effect on carbon and nitrogen decomposition by
altering soil microbial abundance and community function (Zabowski et al. 2008; Wu et al. 2019).
Despite the fact that our initial hypothesis was not fully supported, the findings still provide an
evidence for optimizing land use in implementing a silvopasture system. As observed not only in
this study but by others (Feliciano et al. 2018, Jose et al. 2019), the successful adoption of a
silvopasture system through strip thinning has a potential to sequester more soil carbon while
enhancing species biodiversity, through a deliberate integration of forage as forage can also
sequester carbon not only belowground but aboveground as well.

79
LITERATURE CITED
Axelson, J.N., Alfaro, R.I., Hawkes, B.C., 2009. Influence of fire and mountain pine beetle on
the dynamics of lodgepole pine stands in British Columbia, Canada. For. Ecol. Manage.
257, 1874–1882. doi:10.1016/j.foreco.2009.01.047
Bai, S.H., Dempsey, R., Reverchon, F., Blumfield, T.J., Ryan, S., Cernusak, L.A., 2017. Effects
of forest thinning on soil-plant carbon and nitrogen dynamics. Plant Soil 411, 437–449.
doi:10.1007/s11104-016-3052-5
Bhattarai, M., Secchi, S., Schoof, J., 2016. An Analysis of the Climate Change Mitigation
Potential through Soil Organic Carbon Sequestration in a Corn Belt Watershed 77–86.
doi:10.1007/s00267-016-0771-6
Blaisdell, R., Conant, R., Dick, W., Dobermann, A., Izaurralde, C., Ransom, M., Rice, C.,
Robertson, P., Stuth, J., Thompson, M., 2003. Recommended Procedures for Collecting ,
Processing , and Analyzing Soil Samples in CASMGS Research Plots Selecting a Plot
Location.
Brevik, E., Fenton, T., Moran, L., 2002. Effect of soil compaction on organic carbon amounts
and distribution, South-Central Iowa. Environ. Pollut. 116, 137–141. doi:10.1016/S02697491(01)00266-4
Cheng, Y., Wang, J., Mary, B., Zhang, J. bo, Cai, Z. cong, Chang, S.X., 2013. Soil pH has
contrasting effects on gross and net nitrogen mineralizations in adjacent forest and grassland
soils in central Alberta, Canada. Soil Biol. Biochem. 57, 848–857.
doi:10.1016/j.soilbio.2012.08.021
Conant, R.T., Paustian, K., 2002. Potential soil carbon sequestration in overgrazed grassland
ecosystems. Global Biogeochem. Cycles 16, 90-1-90–9. doi:10.1029/2001gb001661
Conover, W.J., Johnson, M.E., Johnson, M.M., 1981. A comparative Study of Tests for
Homogeneity of Variances, with Applications to the Outer Continental Shelf Bidding Data.
Technometrics 23, 315–361.

80
Deljouei, A., Sadeghi, S.M.M., Abdi, E., Bernhardt-Römermann, M., Pascoe, E.L., Marcantonio,
M., 2018. The impact of road disturbance on vegetation and soil properties in a beech stand,
Hyrcanian forest. Eur. J. For. Res. 137, 759–770. doi:10.1007/s10342-018-1138-8
Dingaan, M.N. V, Tsubo, M., Walker, S., Newby, T., 2017. Soil chemical properties and plant
species diversity along a rainfall gradient in semi-arid grassland of South Africa. Plant Ecol.
Evol. 150, 35–44.
Feliciano, D., Ledo, A., Hillier, J., Nayak, D.R., 2018. Which agroforestry options give the
greatest soil and above ground carbon benefits in different world regions? Agric. Ecosyst.
Environ. 254, 117–129. doi:10.1016/j.agee.2017.11.032
Finzi, A.C., Breemen, N. Van, Canham, C.D., 1998. Canopy Tree-Soil Interactions within
Temperate Forests: Species Effects on Soil Carbon and Nitrogen. Ecol. Appl. 8, 440.
doi:10.2307/2641083
Fischer, P.W., Cullen, A.C., Ettl, G.J., 2017. The Effect of Forest Management Strategy on
Carbon Storage and Revenue in Western Washington: A Probabilistic Simulation of
Tradeoffs. Soc. Risk Anal. 37, 173–192. doi:10.1111/risa.12611
Frank, A.B., Tanaka, D.L., Hofmann, L., Follett, R.F., 1995. Soil carbon and nitrogen of
Northern Great Plains grasslands as influenced by long-term grazing. J. Range Manag. 48,
470–474. doi:10.2307/4002255
Grady, K.C., Hart, S.C., 2006. Influences of thinning, prescribed burning, and wildfire on soil
processes and properties in southwestern ponderosa pine forests: A retrospective study. For.
Ecol. Manage. 234, 123–135. doi:10.1016/j.foreco.2006.06.031
Gurung, B., Nelson, K.C., Smith, J.L.D., 2009. Impact of grazing restrictions on livestock
composition and husbandry practices in Madi Valley, Chitwan National Park, Nepal.
Environ. Conserv. 36, 338–347. doi:10.1017/S0376892910000160
Hofmeister, J., Hošek, J., Brabec, M., Hédl, R., Modrý, M., 2013. Strong influence of longdistance edge effect on herb-layer vegetation in forest fragments in an agricultural
landscape. Perspect. Plant Ecol. Evol. Syst. 15, 293–303. doi:10.1016/j.ppees.2013.08.004

81
Hood, S.M., Baker, S., Sala, A., 2016. Fortifying the forest: thinning and burning increase
resistance to a bark beetle outbreak and promote forest resilience. Ecol. Appl. 26, 1984–
2000. doi:10.1002/eap.1363
Howlett, D.S., Moreno, G., Mosquera Losada, M.R., Nair, P.K.R., Nair, V.D., 2011a. Soil
carbon storage as influenced by tree cover in the Dehesa cork oak silvopasture of centralwestern Spain. J. Environ. Monit. 13, 1897–1904. doi:10.1039/c1em10059a
Howlett, D.S., Mosquera-Losada, M.R., Nair, P.K.R., Nair, V.D., Rigueiro-Rodríguez, A.,
2011b. Soil Carbon Storage in Silvopastoral Systems and a Treeless Pasture in
Northwestern Spain. J. Environ. Qual. 40, 825–832. doi:10.2134/jeq2010.0145
Jose, S., Walter, D., Mohan Kumar, B., 2019. Ecological considerations in sustainable
silvopasture design and management. Agrofor. Syst. 93, 317–331. doi:10.1007/s10457-0160065-2
Kemmitt, S.J., Wright, D., Goulding, K.W.T., Jones, D.L., 2006. pH regulation of carbon and
nitrogen dynamics in two agricultural soils. Soil Biol. Biochem. 38, 898–911.
doi:10.1016/j.soilbio.2005.08.006
Kneeshaw, D.D., Leduc, A., Drapeau, P., Gauthier, S., Pare, D., Carignan, R., Doucet, R.,
Bouthillier, L., Messier, C., 2000. Development of integrated ecological standards of
sustainable forest management at an operational scale. For. Chron. 76, 481–493.
doi:10.5558/tfc76481-3
Krotz, L., Leone, F., Giazzi, G., 2016. High Accuracy of Nitrogen , Carbon and Sulfur Analysis
for Agronomy Applications using the Thermo Scientific Flash Smart Elemental Analyzer,
Thermo Fisher Scientific.
Kurz, W.A., Dymond, C.C., Stinson, G., Rampley, G.J., Neilson, E.T., Carroll, A.L., Ebata, T.,
Safranyik, L., 2008. Mountain pine beetle and forest carbon feedback to climate change.
Nature 452, 987–990. doi:10.1038/nature06777
Lal, R., Griffin, M., Apt, J., Lave, L., Morgan, M.G., 2004. Managing Soil Carbon. Science (80-.
). 304, 393. doi:10.1126/science.1093079

82
Lavelle, P., Decaëns, T., Aubert, M., Barot, S., Blouin, M., Bureau, F., Margerie, P., Mora, P.,
Rossi, J.P., 2006. Soil invertebrates and ecosystem services. Eur. J. Soil Biol. 42, 1–15.
doi:10.1016/j.ejsobi.2006.10.002
Li-COR Biosciences, 2010a. Li-8100 Automated Soil CO2 Flux System & LI-8150 Multiplexer.
Theory of operation. Instruction Manual. Lincoln, NE 68504 USA.
Li-COR Biosciences, 2010b. SoilFluxPro Software. Instruction Manual. 4647 Superior Street.
Lincoln,Nebraska 68504.
Luo, Y., Su, B., Currie, W.S., Dukes, J.S., Finzi, A., Hartwig, U., Hungate, B., MCmurtrie, R.E.,
Oren, R., Parton, W.J., Pataki, D.E., Shaw, M.R., Zak, D.R., Field, C.B., 2004. Progressive
Nitrogen Limitation of Ecosystem Responses to Rising Atmospheric Carbon Dioxide.
Bioscience 54, 731. doi:10.1641/0006-3568(2004)054[0731:pnloer]2.0.co;2
Madsen, R., Xu, L., Claassen, B., McDermitt, D., 2009. Surface Monitoring Method for Carbon
Capture and Storage Projects. Energy Procedia 1, 2161–2168.
doi:10.1016/j.egypro.2009.01.281
Maynard, D.G., Curran, M.P., 2006. Soil Sampling and Methods of Analysis: Bulk Density
Measurement in Forest Soils, Canadian Society of Soil Science.
Melillo, J.M., Aber, J.D., Linkins, A.E., Ricca, A., Fry, B., Nadelhoffer, K.J., 1989. Carbon and
nitrogen dynamics along the decay continuum: Plant litter to soil organic matter. Plant Soil
115, 189–198. doi:10.1007/BF02202587
Mohd Razali, N., Bee Wah, Y., 2011. Power comparisons of Shapiro-Wilk, KolmogorovSmirnov, Lilliefors and Anderson-Darling tests. J. Stat. Model. Anal. 2, 21–33.
Novoa, R., Loomis, R.S., 1981. Nitrogen and Plant Production. Plant Soil 58, 177–204.
doi:0032-079X/81/0582-0177504.20
Olajuyigbe, S., Tobin, B., Saunders, M., Nieuwenhuis, M., 2012. Forest thinning and soil
respiration in a Sitka spruce forest in Ireland. Agric. For. Meteorol. 157, 86–95.
doi:10.1016/j.agrformet.2012.01.016

83
Picchio, R., Neri, F., Petrini, E., Verani, S., Marchi, E., Certini, G., 2012. Machinery-induced
soil compaction in thinning two pine stands in central Italy. For. Ecol. Manage. 285, 38–43.
doi:10.1016/j.foreco.2012.08.008
Pineiro, G., Paruelo, J.M., Oesterheld, M., Jobbágy, E.G., 2010. Pathways of grazing effects on
soil organic carbon and nitrogen. Rangel. Ecol. Manag. 63, 109–119. doi:10.2111/08-255.1
Post, W.M., Kwon, K.C., 2000. Soil carbon sequestration and land-use change: processes and
potential. Glob. Chang. Biol. 6, 317–327. doi:https://doi.org/10.1046/j.13652486.2000.00308.x
Puget, P., Lal, R., 2005. Soil organic carbon and nitrogen in a Mollisol in central Ohio as
affected by tillage and land use. Soil Tillage Res. 80, 201–213.
doi:10.1016/j.still.2004.03.018
Rab, M.A., 1994. Changes in physical properties of a soil associated with logging of Eucalyptus
regnan forest in southeastern Australia. For. Ecol. Manage. 70, 215–229. doi:10.1016/03781127(94)90088-4
Rice, W.A., Penney, D.., Nyborg, M., 1977. Effects of soil acidity on Rhizobia numbers,
nodulation and nitrogen fixation by alfalfa and red clover. Can. J. Soil. Sci 57, 197–203.
Richardson, H.J., Hill, D.J., Denesiuk, D.R., Fraser, L.H., 2017. A comparison of geographic
datasets and field measurements to model soil carbon using random forests and stepwise
regressions (British Columbia, Canada). GIScience Remote Sens. 54, 573–591.
doi:10.1080/15481603.2017.1302181
Saunders, M., Tobin, B., Black, K., Gioria, M., Nieuwenhuis, M., Osborne, B.A., 2012. Thinning
effects on the net ecosystem carbon exchange of a Sitka spruce forest are temperaturedependent. Agric. For. Meteorol. 157, 1–10. doi:10.1016/j.agrformet.2012.01.008
Scurlock, J.M.O., Hall, D.O., 1998. The global carbon sink: a grassland perspective. Glob.
Chang. Biol. 4, 229–233. doi:10.1046/j.1365-2486.1998.00151.x
Sharrow, S.H., 2007. Soil compaction by grazing livestock in silvopastures as evidenced by

84
changes in soil physical properties. Agrofor. Syst. 71, 215–223. doi:10.1007/s10457-0079083-4
Shrestha, R.K., Alavalapati, J.R.R., 2004. Valuing environmental benefits of silvopasture
practice: A case study of the Lake Okeechobee watershed in Florida. Ecol. Econ. 49, 349–
359. doi:10.1016/j.ecolecon.2004.01.015
Silveira, M.L., Xu, S., Adewopo, J., Franzluebbers, A.J., Buonadio, G., 2014. Grazing land
intensification effects on soil C dynamics in aggregate size fractions of a Spodosol.
Geoderma 230–231, 185–193. doi:10.1016/j.geoderma.2014.04.012
Smith, P., Cotrufo, M.F., Rumpel, C., Paustian, K., Kuikman, P.J., Elliott, J.A., McDowell, R.,
Griffiths, R.I., Asakawa, S., Bustamante, M., House, J.I., Sobocká, J., Harper, R., Pan, G.,
West, P.C., Gerber, J.S., Clark, J.M., Adhya, T., Scholes, R.J., Scholes, M.C., 2015.
Biogeochemical cycles and biodiversity as key drivers of ecosystem services provided by
soils. Soil 1, 665–685. doi:10.5194/soil-1-665-2015
Sokolov, A.P., Kicklighter, D.W., Melillo, J.M., Felzer, B.S., Schlosser, C.A., Cronin, T.W.,
2008. Consequences of considering carbon-nitrogen interactions on the feedbacks between
climate and the terrestrial carbon cycle. J. Clim. 21, 3776–3796.
doi:10.1175/2008JCLI2038.1
Tang, J., Qi, Y., Xu, M., Misson, L., Goldstein, A.H., 2005. Forest thinning and soil respiration
in a ponderosa pine plantation in the Sierra Nevada. Tree Physiol. 25, 57–66.
doi:10.1093/treephys/25.1.57
Thomas, S.C., Halpern, C.B., Falk, D.A., Liguori, D.A., Austin, K.A., 1999. Plant diversity in
managed forests: Understory responses to thinning and fertilization. Ecol. Appl. 9, 864–879.
doi:10.1890/1051-0761(1999)009[0864:PDIMFU]2.0.CO;2
Turner, J., Lambert, M., 2000. Change in organic carbon in forest plantation soils in eastern
Australia. For. Ecol. Manage. 133, 231–247. doi:10.1016/S0378-1127(99)00236-4
Vesterdal, L., Raulund-Rasmussen, K., 1998. Forest floor chemistry under seven tree species
along a soil fertility gradient. Can. J. For. Res. 28, 1636–1647. doi:10.1139/x98-140

85
Wang, Q., Li, Y., Wang, Y., 2011. Optimizing the weight loss-on-ignition methodology to
quantify organic and carbonate carbon of sediments from diverse sources. Environ. Monit.
Assess. 174, 241–257. doi:10.1007/s10661-010-1454-z
Wang, X., Wang, J., Zhang, J., 2012. Comparisons of Three Methods for Organic and Inorganic
Carbon in Calcareous Soils of Northwestern China. PLoS One 7.
doi:10.1371/journal.pone.0044334
Wic Baena, C., Andrés-Abellán, M., Lucas-Borja, M.E., Martínez-García, E., García-Morote,
F.A., Rubio, E., López-Serrano, F.R., 2013. Thinning and recovery effects on soil properties
in two sites of a Mediterranean forest, in Cuenca Mountain (South-eastern of Spain). For.
Ecol. Manage. 308, 223–230. doi:10.1016/j.foreco.2013.06.065
Wu, R., Cheng, X., Han, H., 2019. The effect of forest thinning on soil microbial community
structure and function. Forests 10. doi:10.3390/f10040352
Zabowski, D., Chambreau, D., Rotramel, N., Thies, W.G., 2008. Long-term effects of stump
removal to control root rot on forest soil bulk density, soil carbon and nitrogen content. For.
Ecol. Manage. 255, 720–727. doi:10.1016/j.foreco.2007.09.046
Zhang, Y., Biswas, A., Adamchuk, V.I., 2017. Implementation of a sigmoid depth function to
describe change of soil pH with depth. Geoderma 289, 1–10.
doi:10.1016/j.geoderma.2016.11.022

86

CHAPTER 4 – MANAGEMENT IMPLICATIONS AND FUTURE
RESEARCH
Forest thinning which is a selective removal of some trees, provides an opportunity to improve
the growth and health of the remaining trees or enhance understory vegetation (Verschuyl et al.
2011). Forest thinning is a crucial practice in forest management, as it can be carried out to reduce
forest fire fuel loads and can be a factor to slowdown the spread of tree diseases (Smirnova et al.
2021). In addition, the forest thinning approaches applied in the Goudie lodgepole pine forest
focused on integrating ranching and forest industries using strip thinning, so that both forest and
range products can be fully realized (Appendix- A.14). However, thinning involves the use of
heavy machinery and equipment which often results in land disturbance (Picchio et al. 2012). Soil
compaction and litter clearing exposes the soil, which can alter soil properties, hydrology and
aboveground species development (Kozlowski 1999). It is important to understand all the factors
involved and the subsequent benefits in order to make informed and sustainable forest management
decisions.
My study assessed the influence of different forest thinning treatments using strips harvesting
methods on forage quality and quantity as well as soil carbon storage. The strip-harvested areas
were compared to un-thinned areas in order to contribute to a wider understanding of the benefits
and effects of forest thinning in British Columbia forested rangeland sites.
Based on our investigation and results, the 20 m and 15 m thinned treatments seemed to
influence the most forage growth for livestock and provided higher diverse and richness of
understory plant species. Both thinned and un-thinned areas seemed to produce a comparable
amount of organic matter, carbon storage and carbon flux. Although the findings of my three years
study provided an insight on the potential of improving forage production, and carbon
sequestration in a silvopasture system, further research assessing long-term impacts on carbon
sequestration and carbon flux in thinned versus un-thinned areas is necessary to determine the full
potential of this type of forest management approach.
A long-term experiment should be conducted as well to evaluate at what extent soil bulk density
influences forage productivity and carbon sequestration in thinned forested areas and
understanding the impact of livestock grazing on forage regeneration and soil carbon.
In addition, it is important to conduct a long-term study on the growth and health response of
the remained trees post timber harvesting.

87
LITERATURE CITED
Kozlowski, T.T., 1999. Soil Compaction and Growth of Woody Plants. Scand. J. For. Res. 14,
596–619. doi:10.1080/02827589950154087
Picchio, R., Neri, F., Petrini, E., Verani, S., Marchi, E., Certini, G., 2012. Machinery-induced
soil compaction in thinning two pine stands in central Italy. For. Ecol. Manage. 285, 38–43.
doi:10.1016/j.foreco.2012.08.008
Smirnova, E., Bennett, P., Egan, J., Kalachev, L., Goodburn, J., Lockman, I.B., Hartless, C.,
2021. Impacts of thinning treatments on dynamics of needle disease caused by Elytroderma
deformans (Weir) Darker and interactions with bark beetle-attacks in the northern Rocky
Mountains. For. Ecol. Manage. 480. doi:10.1016/j.foreco.2020.118654
Verschuyl, J., Riffell, S., Miller, D., Wigley, T.B., 2011. Biodiversity response to intensive
biomass production from forest thinning in North American forests - A meta-analysis. For.
Ecol. Manage. 261, 221–232. doi:10.1016/j.foreco.2010.10.010

88
APPENDIX – A
Relationships between understory species community in the thinned areas Pre-harvest (July
2018) and Post-harvest (August 2019 and July 2020).

Figure A.1. PCoA analysis based on Bray-Curtis dissimilarity for July 2018 understory species
community composition between areas assigned for 10 m and 15 m treatment widths.

Figure A.2. PCoA analysis based on Bray-Curtis dissimilarity for July 2018 understory species
community composition between areas assigned for 10 m and 20 m treatment widths.

89

Figure A.3. PCoA analysis based on Bray-Curtis dissimilarity for July 2018 understory species
community composition between areas assigned for 15 m and 20 m treatment widths.

Figure A.4. PCoA analysis based on Jaccard dissimilarity for August 2019 understory species
community composition between 10 m and 15 m thinned treatments.

90

Figure A.5. PCoA analysis based on Jaccard dissimilarity for August 2019 understory species
community composition between 10 m and 20 m thinned treatments.

Figure A.6. PCoA analysis based on Jaccard dissimilarity for August 2019 understory species
community composition between 15 m and 20 m thinned treatments.

91

0.75

psuedoF = 2.34;
p = 0.001

PCoA2

0.50
0.25

Treatments

0.00

10 m
15 m

-0.25
-0.50
-0.6

-0.3

0.0
PCoA1

0.3

0.6

Figure A.7. PCoA analysis based on Jaccard dissimilarity for July 2020 understory species
community composition between 10 m and 15 m thinned treatments.
psuedoF = 3.99;
p = 0.001

PCoA2

0.4
Treatments
0.0

10 m
20 m

-0.4

-0.4

0.0
0.4
PCoA1
Figure A.8. PCoA analysis based on Jaccard dissimilarity for July 2020 understory species

community composition between 10 m and 20 m thinned treatments.

92
psuedoF = 1.7;
p = 0.009

PCoA2

0.4
Treatments
15 m
20 m

0.0

-0.4

-0.4

0.0
PCoA1

0.4

Figure A.9. PCoA analysis based on Jaccard dissimilarity for July 2020 understory species
community composition between 15 m and 20 m thinned treatments.

A.10. Soil Horizons, texture and coarse fragment determination pre-timber harvest in three
adjacent block design.

93

A.11. Soil texture diagram of Goudie silvopasture research area contains 46.4% sand, 47.9% silt
and 5.8% clay.

94

A.12. Excavation method and materials for soil bulk density analysis used post-harvest 2020.

A.13. An NDVI map example that shows soil compaction in the strips post timber harvesting.

95

A.14. Integrated forage, livestock and timber approach through forest strip thinning.

96

A.15: Map of one of the experimental design blocks detailing the strip widths and their timber
buffers.

97
A.16: Plant species identified from all three blocks in Goudie, Kelowna Pre and post strip
thinning. (USDA- United States Department of Agriculture 2019)
Seeded agronomic species
Common name

Scientific name

Orchard grass
Meadow Brome

Dactylis
glomerata L.
Bromus riparius

Intermediate
Wheatgrass
White clover

Thinopyrum
intermedium
Trifolium repens

Unseeded palatable
Common
name
Kentucky
Bluegrass
Idaho fescue

Scientific name

Canada
Bluejoint
Western
fescue
Pinegrass

Calamagrostis canadensis

Unseeded non-palatable species
Common
name
Graceful
cinquefoil
Common
snowberry
Soopolallie

Scientific name

Spiraea betulifolia

Calamagrostis rubescens

Birch-leaved
spirea
Falsebox

Smooth
Brome
Sedges

Bromus inermis

Sitka Alder

Alnus viridis

Carex spp.

Twinflower

Linnaea borealis

Fireweed

Chamaenerion angustifolium

Bunchberry

Cornus canadensis

Dandelion

Taraxacum erythrospermum

One-leaved
foamflower
Sweet-scented
bedstraw
Round leaf
viola
Coltsfoot

Tiarella
trifoliata var. unifoliata
Galium triflorum

Prince's pine

Chimaphila umbellata

Utah
honeysuckle
Field
chickweed
False
solomon's-seal
Hawkweed

Lonicera utahensis

Tiger lily

Lilium lancifolium

Common daisy
spp.
Pearly
everlasting
Fragile sour
weed/Common
sheep sorrel
Buttercup

Bellis perennis

Thistle

Cirsium spp.

Salsify

Tragopogon porrifolius

Prickly wild
rose
Rattlesnake
plantain

Rosa acicularis

Poa pratensis L.
Festuca idahoensis Elmer

Festuca occidentalis

Potentilla gracilis
Symphoricarpos albus
Shepherdia canadensis

Paxistima myrsinites

Viola rotundifolia
Tussilago farfara

Cerastium arvense
Maianthemum
racemosum
Hieracium spp.

Anaphalis margaritacea
Rumex acetosella

Ranunculus spp.

Goodyera pubescens

98
black
huckleberry
Common
mitrewort
Salmonberry

Gaylussacia baccata

Vaccinum spp

Vaccinum spp

Lodgepole
seedling pine
tree
Columbine

Pinus contorta

Mountain
sweet cicely
Common
Juniper
White avens

Osmorhiza chilensis

One-sided
wintergreen
Western
meadow-rue

Orthilia secunda

Mitella nuda L
Rubus spectabilis

Aquilegia spp.

Juniperus communis
Geum canadense

Thalictrum occidentale