THE EFFECT OF NATURAL FENCEROWS ON LOCAL STANDARDIZED W'INDSPEED, TEMPERATURE AND RELATIVE

H[JIMIDZTY

Million Marno Bayou

A thesis submitted in conformity with the requirements for the degree of Master of Science in Forestry

Graduate Department of Forestry University of Toronto

O Copyright by Million Mamo Bayou 1997

National Library 141 of Canada Bibliothèque nationale du Canada

Acquisitions and Acquisitions et Bibliographie Services services bibliographiques

395 Wellington Street 395. nie Wellington Ottawa ON K1AON4 OttawaON K1AON4 Canada Canada

The author has granted a non- exclusive licence allowing the National Library of Canada to reproduce, loan, distribute or sell copies of this thesis in microfom, paper or electronic formats.

The author retains ownership of the copyright in this thesis. Neither the thesis nor substantial extracts f h n it may be printed or otherwise reproduced without the author's permission.

L'auteur a accorde une licence non exclusive permettant à la Bibliothèque nationale du Canada de reproduire, prêter, distribuer ou vendre des copies de cette thèse sous la forme de rnicrofiche/film, de reproduction sur papier ou sur format électronique.

L'auteur conserve la propriété du droit d'auteur qui protège cette thèse. Ni la thèse ni des extraits substantiels de celle-ci ne doivent être imprimés ou autrement reproduits sans son autorisation.

THE EFFECT OF NATURAL FENCEROW ON LOCAL STANDARDUED WINDSPEED, TEMPERA12TRE AND RELATLVE HUMlDITY

Million Mamo Bayou M.Sc.F. 1997 Graduate Department of Forestry University of Toronto

ABSTRACT

The effect of natural fencerows on the local microclimzte was studied at two different site

at Lunenberg in 1995 and at Metcalfe (Eastern Ontario) 1996. Windspeed, air

temperature and relative hurnidity were measured at six stations perpendicular to the

fencerow. Data for each of the five stations closest to the fencerows were standardized

relative to the station fmhest downwind (beyond a signifiant influence of the fencerow).

The porosity of the fencerowg an important factor in determining the efficiency, was

determined by using photographs and GIS sofi ware program.

Standardized windspeed and relative hurnidity did not differ significantly with distance

fiom the fencerow at the Lunenberg site. At Metcalfe standardized windspeed,

temperature and relative humidity were significantiy afKected 'ay the variation in distances

From the natural fencerows. Porosity at the two fencerows was sirnilar, 30 and 33 % at

Lunenberg and Metcalfe respectively, but the coefficient of variation (CV) was higher at

Lunenberg , 0.45 vs 0.23, indicating greater heterogeneity.

ACKNOWLEDGMENTS

1 would like to sincerely acknowledge and express my gratitude to my supervisor Dr. W.

Andrew Kenney for giving me the opportunity to pursue this area of study, and for his

guidance throughout the leaming process. This challenging project would not have been

a reality without his patience and advice. 1 wish to express my deepest thanks to the

members of my advisory committee, Dr. D. Fayle and Dr. T. Blake, for their valuable

advice and criticism. My sincere thanks is also extended to DonMcIvor from

Environment Canada, for his valuable comrnents and material support.

The financial support of Eastern Ontario Mode1 Forest for this project was very

instrumental in the completion of this document, and their support is duly acknowledged.

1 am in debt to the landowners, Mr. Dalton Adams and Mrs. Helen Dike, for allowing me

to use their property and their hospitality dunng my stay in Eastern Ontario. The

author wish to thank Dr. Alan Darlington, T. H. Tsegenet, Tsegaye W. Mariam, M.

Ngusya, and Daniela Punc-Mladenovic for their respective contributions at various

stages of the project.

Finally, al1 my families members, particularly my younger brother, Yared M. Bayou,

deserve my appreciation and recognition for their continued encouragement and financial

support.

TABLE OF CONTENT

Content Pane

ABSTRACT

ACKNOWLEDGMENTS

TABLE OF CONTENT

LIST OF TABLE

LIST OF FIGURES

LIST OF APPENDICES

LIST OF ABBREVIATIONS

FNTRODUCTION

Background

The Research problem

The study objectives and hypothesis

LITERATURE REVIEW

Some objectives of shelter

Natural fencerows

Effect of shelter on microclimate

Windbreak design

Effect of shelter on crop yields

Vegetables

Field crops

. . 11

.-. 111

iv

vi

vii

viii

X

1

1

4

5

6

6

7

9

13

16

17

18

MATERIALS AND METHODS

Site selection

Data collection

Porosity estimation

Data analysis

RESULTS

Standardized windspeed (SWS)

Standardized temperature (ST)

Standardized relative humidity (SRH)

Porosity Estimation

DISCUSSION AND CONCLUSION

Microclimate Effect

Physical Structure (Porosity)

CONCLUSION

RESEARCH NEEDS

LITERATURE CITED

APPENDICES

LIST OF TABLES

Tables Pane

1. Expected means squares for each source for standardizec

windspeed, temperature and relative humidity used

to test the hypothesis

2. Estimated porosity of nine plots along the fencerow

at the Lunenberg site

3 . Estimated porosity of seven plots along the fencerow

at the Metcalfe site

LIST OF FIGURES

Figure

Location map indicating the Lunenberg and Metclafe sites

Aerial photograph of Lunenberg site.

Aerial photograph of Metcalfe site.

Standardized windspeed (SWS) profile as affecte( i by the distance

(stations) from the natural fencerow at Lunenberg site

Standardized windspeed (SWS) profile as afEected by the distance

(stations) fiom the naturai fencerow at Metcalfe site

Standardized temperature (ST) profile as af5ected by the distance

(stations) From the natural fencerow at Lunenberg site

Standardized temperature (ST) profile as affected by the distance

(stations) frorn the natural fencerow at Metcalfe site

Standardized relative humidity (SRS) profile as affected by the

distance (stations) fiom the natural fencerow at Lunenberg site

Standardized relative humidity (SRH) profile as afTected by the

distance (stations) ftom the natural fencerow at Metcalfe site

Digitked photograph of the Lunenberg fencerow used in the estimation

of nine sample plots dong the natural fencerow 36

Digitized photograph of the Metcalfe fencerow used in the estimation

of seven sample plots dong the natural fencerow 37

vii

LIST OF APPENDICES

Appendices Page

Photographs of Lunenberg fencerow

Photographs of Metcalfe fencerow

The Environment Canada monthiy meteorological surnmary of Ottawa

Airport for the months of July, August and September for 1995

and 1996 52

A sumrnary of the ANOVA used to test the hypothesis (Ho) of no

significant difference among days, hour, stations and their interactions with

respect to standardized windspeed (SWS) for Lunenberg Fencerow Site 53

A surnmary of the ANOVA used to test the hypothesis (Ho) of no significant

difference among days, hour, stations and their interactions with respect to

standardized windspeed (SWS) for Metcalfe Fencerow Site 54

A surnmary of the NOVA used to test the hypothesis (Ho) of no significant

difference arnong days, hour, stations and their interactions with respect to

standardized temperature (ST) for Lunenberg Fencerow Site 55

Standardized temperature (ST) as affected by the interaction of staion and

day for the natural fencerow at Lunenberg Site. 56

A summary of the ANOVA used to test the hypothesis (Ho) of no significant

difference arnong days, hour, stations and their interactions with respect to

standardized tem~erature (ST) for Metcalfe Fencerow Site 57

viii

9. A summary of the ANOVA used to test the hypothesis (Ho) of no significant

difference among days, hour, stations and their interactions with respect to

standardized relative hurnidity (SRH) for Lunenberg Fencerow Site 58

10. A sumrnary of the ANOVA used to test the hypothesis (Ho) of no significant

difference among days, hour, stations and their interactions with respect to

standardized relative hurnidity (SRH) for Metcaife Fencerow Site 59

LIST OF ABBREVIATION

EMS

ANOVA

GIS

cv

C.E.

%

sws

ST

SRH

Expected means square

Analysis of variance

Geographical information system

coefficient of variations

coefficient of error

percent

standardized windspeed

standardized temperature

standardized relative humidity

INTRODUCTION

Background

The recognition of wind effects and the need for wind protection in order to improve crop

growth, livestock productivity, and human welfare dates back to the 15" century in

Scotland (Cabom, 1965). In North Arnerica, and in particular in the Great Plains of the

United States and the Canadian Prairies, severe crop losses and soi1 erosion problems in

the 1930s resulted in the planting of windbreaks on a large scale (Staple and Lehane,

1955; Rosenberg, 1983).

The primary function of tree windbreaks is to reduce and modiQ local wind rnovement.

This modification in turn affects other environmental factors that have an impact on plants

and anirnals. Numerous studies on the aerodynamics of shelter reported that wind velocity

decreases up to certain distances behind various windbreaks (Gloyne, 1954; Grace, 1977;

Mulheam and Bradey, 1977; McNaughton, 1988). In the lee of windbreaks, windspeed is

reduced with increasing distance fkom the barrier, up to a certain distance. M e r this

there is a gradua1 increase in velocity. This recovery of windspeed is attributed to factors

such as porosity of the windbreak, slope of the terrain, surface roughness, the angle of

incidence of the wind, and windspeed (Cabom, 1965; Hagen and Skidmore, 197 1 ; K e ~ e y

1986).

The complex relationships between vanous microclimate parameters make it difficult to

fully understand the eEect of shelter on its local environment. Nevertheless, wind

influences the microclimate of plants through its direct effect on principal meteorological

elements such as temperature, relative humidity, and evaporation (Rosenberg, 1983;

McNaughton, 1988). Similarly, another indirect effect of windbreaks is on soi1 moisture.

Shelterbelts influence soi1 rnoisture by controlling the snow cover during the winter

season, and by affecting the evaporation of moisture from the soi1 surface (Staple and

Lehane, 1955; Frank et ai., 1974a; Peterson, 1984).

This study was conducted on two sites in Eastern Ontario (Figure l), and focused on

fencerows, that are naturally established live fences usually composed of hardwoods trees,

shmbs, and grasses. In contrast to planted field shelterbelts or windbreaks, that are

established for the purpose of protection from the wind, natural fencerows usually have

been lefl to delineate property lines, or to keep cattle out of croplands.

The orientation and stmcture of these natural bamers might not be optimal for use as field

windbreaks. However, they do provide some level of protection from the adverse effects

of wind on adjacent fields. A well managed fenceraw may give al1 the benefits of

shelterbelts or windbreaks without the cost associated with their establishment and the

longer tirne needed for planted windbreaks to grow. Furthemore, because of the

reiatively undisturbed vegetation and the litter conditions fencerows provide, they c m be

an important habitat to non-game wildlife in agricultural areas.

In this document a shelterbelt is defined as a belt of trees andlor shmbs that is usually

more than two rows wide, and arranged as a protection for fields and crops against strong

winds (van Eimem, et ai-. 1962; Kenney, 1986). The term windbreak applies to short

barriers of one or two rows of trees designed to obstmct wind flow and intended to

Figure 1. Location map indicating the Lunenberg and Metcalfe study sites.

protect buildings, soil, crops, or livestock from the effect of winds (Kemey, 1986).

Shelter extent is the distance to which a significant reduction in windspeed extends, and

usually is measured in multiples of the average height (H) of the wind bamer (Cabom,

1957; K e ~ e y , 1986). Shelter quality is the magnitude of reduction in wind velocity

expressed as a percentage of the velocity in an adjacent unsheltered area (Kemey, 1986).

The Research Probtem

The trend toward removing natural fencerows has increased with agriculturai

development. It is largely attnbuted to the belief that they take up valuable cropland, and

provide little or no benefit. This trend rnay be seriously affecting the level of biodiversity,

wildlife habitat, and the productivity of farmland. If farrners can be s h o w that fencerows

provide an economic benefit through improved yield, it may be easier to convince them to

conserve fencerows, and thereby contribute to increased biodiversity and habitat.

Windbreaks have been shown to have a direct impact on microciimate factors such as

windspeed, temperature, and relative hurnidity. Since these parameters have been s h o w

to have an impact on crop yield and quality, they were considered to be good indicators of

the impact of fencerows on crop yield.

No systematic study of the impact of fencerows on the adjacent microclimate has been

conducted in the province. Nor I am aware of any such work in the Canadian context.

Little data exists on natural fencerows apart from studies of wildlik movements in these

habitats (Wegner and Memam, 1979; Best, 1983; Bennett et al., 1994). Our Iack of

knowledge rnay have contributed to the failure to recognize their value as wind barriers,

and to the decrease of fencerow coverage within the agricultural landscape. In this

respect, it is essential to study the effects of fencerows in the Canadian context.

This project therefore address the effect of fencerows on these microclimate parameters.

Furthemore, considering the significant effect of porosity on the efficiency of planted

windbreraks, and the fact that natural fencerows exhibits irregular physical structure, an

assessrnent of porosity is incorporated in the study.

It is beyond the scope of this study to quanti@ the effect of vegetation type, distribution,

the effects of fencerows on wind erosion, or this impacts on hurnan and animal wel1 being.

Since actual crop yield can be affected by many parameten (soi1 type, fertility, genetics,

plant density, etc.) it would be difficult to separate these factors from the impact of

shelter.

The Study Objectives and Hypothesis

The overall objective of this research is to study role of natural as field windbreaks. To

address this general objective, the following specific objective and nul1 hypothesis (FI,,)

were designed.

Objective: To determine the effect of natural fencerows on windspeed, air

temperature, and relative humidity.

There is no significant difference among positions downwind from a naturai

fencerow with respect to windspeed, air temperature, and relative humidity.

LITERATURE REVIEW

Most of the published information dealing with shelter effects relates to windbreaks and

shelterbelts, the following discussion will refer to these artificial structures rather than to

naturd fencerows.

Some Objectives of Shelter

Trees and shmbs are commonly planted to protect plants, animals, and humans from the

effects of wind. Shelter plays a vital role in reducing windspeed and modifying

rnicroclimates to irnprove crop yields and quality (Grace, 1977; Rosenberg, 1983;

Baldwin, 1988; Sun and Dickinson, 1994). Improvements include early ripening of fi-uits

(Bagley, 1964), increased crop quality (Fortin, 1986), and increases yield of various crops

(Frank et al., 1974b; Monnette and Stuart, 1987). Shelter is also responsible for reduced

plant damage due to wind and wind-blown soil. In areas where snow provides a source of

moisture, the even distribution of snow over fields due to windbreaks increases the

probability of higher yields of crops. Windbreaks retard snow movement allowing 80%

to 90% of moisture to soak into the ground for the following crop year (Spanbauer,

1989).

Windbreaks are used to protect Iivestock from severe winter winds, and to provide

beneficial shade in the summer. More than 100 species of birds and 28 species of

mammals have been reported to use windbreaks (Johnson and Beck, 1988). Windbreaks

and fencerows provide benefits to wildlife by protecting them from adverse weather, and

supplying refuge cover, food sites, a reproductive habitat, and travel cot-ridors (Wegner,

and Merriam, 1979; Best, 1983; Bennett, et al., 1994). In addition to the aesthetic value

and to providing secondary forest products, studies in the United States have also proven

that windbreaks can be effective in reducing energy needs for heating and cooling homes

by 10% to 15% (Dewalle and Heisler, 1988).

Natural Fencerows

Historically, in pre-rnechanized agricultural systems, agncultural fields were protected by

tree lines and fencerows that covered a considerable acreage in Nonh Amenca. These

natural bamers were generally left to delineate property boundaries and keep cattle off

cultivated cropland. It should be possible to achieve significant shelter benefits, such as

those outlined above.

Our understanding of the effect of windbreaks on microclimate is far from complete. The

early work of Bates (19 1 1, 1944), Cabom (1957), Van Der Linde (1962), and Jensen

(1954), can be considered as a basis for most microclimate studies on windbreaks, and are

frequently cited. Upon reviewing the literature, one encounters repons on hedgerow

studies, mostly based on European conditions and, in particular, England. There rnight be

some similarity between natural " hedgerows" and " fencerows" . However, fencerow

vegetation in the Canadian context is distinct from the British hedgerows. Hedgerows

were originally cultured, and lower species diversity, a denser structure, are wider, and

are more homogeneous than our North Arnerican natural fencerows (Wegner and

Memam, 1979).

Unfortunateiy, fencerow removal has increased with agricultural development where large

fields were needed to accommodate large tractors and other farm implements. The trend

toward removing fencerows is largely attributed to the opinion that they take up valuable

cropland and provide little or no benefit. It is believed that they reduce yields by

competing with adjacent crops. In addition, sometimes they are also considered to harbor

pests and insects and to damage farm machinery (Berna, undated). Many of these

concems are not supported by research. Little data exists on natural fencerows apart from

studies on wildlife movements in these habitats (Wegner and Merriam, 1979; Best, 1983;

B e ~ e t t , et ai., 1994).

Trees, shmbs, and the associated vegetation of natural fencerows increase biodiversity in

the agricultural landscape. Existing data on fencerows near Ottawa suggest that those

with tall trees and a woodland structure are used as corridors by Chipmunks (7kminr

strialiu lysleri) (Bennett et al., 1994). In another study, 62 direrent species of animals

were observed to be using fencerow habitats. However, it was estimated that fewer than

10 species would be present in the same area in the absence of the fencerow (Best, 1983).

Contrary to the cornmon assumption that fencerows harbor harmfil pests, Spanbauer,

(1989) reported that s h b b y fencerows have a 33% decrease in the number of insects

considered harrnfil, while mammals that feed on insects increased 14 times when

compared to sod fencerows.

Working with planted windbreaks, Baldwin (1988) found that yields were indeed reduced

adjacent to the windbreak, but that the yield gains achieved fùrther away more than offset

these losses in both corn (Zea moys) and soybeans (Glycine m a ) .

Effect of Shelter and Microclimate

In defining the stmctural quality of windbreaks, those with open, medium, and dense

porosity have been recommended depending on the objective (Van Eirnern et al., 1962;

Cabom, 1965). An effective reduction of windspeed by a moderately porous windbreak

to a distance of 15 to 20H has been reported (Kenney, 1986). Likewise, others have

found a substantial reduction of windspeed, from 1OH up to 40H (Gloyne, 1954; Staple

and Lehane 1955; Bagley, 1964; Cabom, 1965; Grace 1977). Kenney (1 986) found that

the porosity of single-row conifer windbreaks was governed more by the species of trees

than by the inter-tree spacing. He suggested that this tendency is similar to that exhibited

in conifer stands. In the latter case, the biomass produced on a given area over time wili

be similar over a relatively large range of planting densities. At high densities, the biomass

will be produced by many smaller trees, whereas in the less dense stand the biomass will

be camed by relatively fewer, but larger trees. Multiple-row windbreaks will have a muîh

lower porosity than single-row windbreaks, since the area of available growing space is

larger (Kemey, 1 986).

Gloyne (1954) argued that the direction of the wind is as important as the windspeed,

because it is more influenced by the topography of the terrain. He emphasized that a

distinction should be made between damaging winds and prevailing winds. He States that:

"the damaging winds for outdoor plants might well be those occurring during any wind

sensitive phases in the cycle of growth and development of the plant". He concluded that

it may not be high windspeed as such that causes damage, but phenomena associated with

the particular wind direction, e.g., sea-salt, low temperatures, or low hurnidity.

Many environmental factors may be created in the microclimate adjacent to windbreaks.

McNaughton (1 988) suggested that there are two microclimate zones. The first is the

"quiet zone" which extends approximately up to 8Y while the second is the "wake zone"

extending beyond 8H. He opines that in the quite zone there is reduced turbulence and

smaller eddy sire. Further downwind, in the wake zone, there will be increased turbulence

with eddy sizes retuming to the upwind scale. Reduced transport of heat, vapour and

carbon dioxide occur in the "quite zone," while increased transport of these environmental

factors occurs in the "wake zone".

The influence of temperature and relative humidity has been neglected in much of the

literature (Rosenberg, 1983). Nevertheless, some of the early works provide reports of

shelter effect on temperature (Woodruffe et al., 1959; Van der Linde, 1962; Guyot, 1963;

Read, 1964; Rosenberg, 1966). They reported that the effect of shelter on air temperature

was closely related to the effect on radiation balance, airflow, and evapotranspiration. The

air temperature was higher on the side facing the Sun, and lower on the shaded side. This

resulted from the interception and reflection of radiation by the shelter (Van der Linde,

1962; Guyot, 1963). Similady, Read (1964) found that during the summer the efEect of

windbreaks in both the windward and leeward directions is more pronounced during

hotter and drier weather, with higher solar radiation, and a greater daily range of

temperature as compared to cooler, wetter weather. Likewise, Rosenberg (1 966)

reported minimal air temperature differences after a sharp drop in radiation.

Generally, long term average data have shown that air temperatures measured near

windbreaks were higher than those measured in the open during the daytime, and lower

during the night (Staple and Lehane, 1955; Woodruffe et al., 1959; Read, 1964).

However, under certain conditions, reduced daytime temperature and increased nighttime

temperature were also observed, or temperatures were simply not affected by the shelter

(Aase and Skiddoway, 1974; Rosenberg, 1966).

Woodruffe et al., (1959) pointed out that the pattern of air temperature to the leeward of

a wind barrier was closely related to the eddy zone created by the bamer. Warm zones air

occurred close to the ground and in proxirnity to trees where predominant eddy current

exits. Observations by Read (1964) corresponded to this theory. He reported a daytime

temperature up to 6" F warmer in the O to 4H leeward zone, and up to 5" F cooler in the 4

to 25 H leeward zone, as compared to open field temperature. In addition, nighttime

temperatures were found to be 3 O F warmer in the O to 25H leeward zone.

Relative humidity has ofien been used to quanti@ atmospheric humidity, because it

indicates both air temperature and humidity conditions. The effect of shelter on relative

humidity is not always unifom. Various factors such as temperature, radiation, soi1

moisture, evaporation, transpiration, and air mixing influence the relative humidity

(Woodruffe et al., 1959; Van der Linde, 1962; Guyot, 1963). Brown and Rosenberg

(1970) reported that the effect of relative humidity depended on the distance from the

shelter. Rosenberg (1 983) observed generally greater relative humidity during the day in

sheltered areas as compared to open fields. Relative humidity in the sheltered zone was

from 2 %to 8% percent higher than in the open (Cabom, 1957).

6' The effects of temperature are partly exerted through water relations because decreasing

temperature is accompanied by decreasing rates of evaporation and transpiration. The

meteorological factors most likely to influence evaporation were found to be temperature,

relative air humidity, and wind speed (Staple and Lehane, 1955). The higher the

windspeed, the greater the evaporation. Similarly, the higher the temperature, and the

lower the relative humidity, the higher the capacity of the air to absorb moisture. For

example, both Miller et ai., (1974) who studied sheitered soybean and wheat fields, and a

review by McNaughton (1988) reported that windbreaks reduce the advection of sensible

heat (horizontal transport of sensible heat energy in the downwind direction) thus reducing

evaporative demand.

Windbreaks influence soil moisture content by controlling snow distribution during the

winter, and by reducing evaporation from the soil. Small reductions in wind velocity

substantially increase snow accumulation dong the bamers (Peterson, 1984). The snow

trapping efficiency of windbreaks is related to the porosity and design of the bamer

(Shaw, 1988). Likewise, snowdnft depth and width is influence by the wind velocity,

topography, and snow availability (Frank et al. 1976b).

It is a matter of controversy in the literature s to the cause and effect of windbreaks on

plant growth. Much of this is due to the lirnited knowledge of the major related

processes: photosynthesis, transpiration, and stomatal conductance. This has made it

difficult to understand the relationships between microclimate and yield (Zhang, 1993).

Despite the apparent contradictions in research results, many workers agree that plant

growth is determined Iargely by water status. Rosenberg (1966) stated that the major

influence of windbreaks on plant growth, particularly under dryland conditions, is due to

the redistribution and conservation of soil water. Water from the snowrnelt, at least in the

temperate zone, is important in improving soil water content, particularly early in the

growing season. Thus a windbreaks influence the accumulation of snow in the winter as

well as the reduction of evapotranspiration during the summer, and thus may have a

positive impact on the growth and development of crops.

Windbreak Design

Although shelter effect are partly a hnction of weather conditions, air mass dominance,

and wind speed (Rosenberg, 1966), studies indicate that the aerial extent, duration, and

magnitude of the shelter effects are controlled by the physical charactenstics of the

windbreak. These charactenstics include its length, height, width, orientation, and

porosity (Cabom, 1965; Rosenberg, 1966; K e ~ e y , 1986). These structurai qualities of

wind bamers are interrelated with height and porosity of the bamer being the most

important factors with respect to its eficiency (Bean et al., 1973; Kenney, 1987; Borrelli,

et al., 1989). Ideally windbreaks are erected at nght angles to the prevailing wind. In the

design of the shelter restricting factors such as the orientation of the property boundanes,

land use, and climatic factors might anse and be beyond the control of the landowner. In

fencerow management, the fact that they are naturally grown and preexisting makes it

impossible to govern their orientation. However, other factors can be controlled once the

objective of the shelter is defined.

The pnmary concern of vanous workers has been to determine the extent of the shelter

produced by the windbreak (Gloyne, 1954; Staple and Lehane 1955; Bagley, 1964; Grace,

1977). Effective shelter protection for distances from 10 times the height (10H) to 40

tirnes the height (40H) from the windbreak have been reported. For example, Cabom

(1965) reported wind reduction for a distance equal to 40 H, of which 30H is on the

leeside of the bamer. Staple and Lehane (1955) showed a wind velocity reduction up to a

distance of 20H.

Broadly speaking, an effective reduction in windspeed (a 20% reduction is considered

significant) is most pronounced on the leeward side of a moderately porous windbreak.

This is normally true to a distance of 15 to 20 times the height Kenney, 1986). An

additional zone of protection exists on the windward side for a distance of 5 to 10 times

the height (Gloyne, 1954; Bagley, 1964). The range of distances up to which significant

reduction is achieved wili be influenced pnmarily by the porosity, and other factors

including topography, windspeed, angle of incidence of the wind relative to the

windbrealg and atmospheric stability.

Apart from the height of shelters, many workers agree that the porosity of a windbreak is

the most important structural factor in the reduction of windspeed (Hagen and Skidmore,

197 1; Bean et al, 1975; Kemey, 1987; Borrelli et ai., 1989; Loeffler et al., 1992). The

porosity of a barrier (which is also referred to as permeability) is expressed as a

percentage of total bamer area, and is a function of the tree species and number of rows

of shelter (Kenney, 1986). Studies indicate that denser barriers exert greater drag and

absorb more momenturn from the mean flow (McNaughton, 1988). Heavily foliated

trees create a solid-wall effect not allowing any wind to pass through the bamer. This

forces the air to move upward and over the top. The largest reduction is close to the lee,

and the largest shear is produced between the retarded air below, and the accelerated air

above. This situation creates turbulence on the leeward side, which could create

unfavorable microclimatic conditions for crops, soil, and animals (McNaughton, 1988).

Porous bamers allow more air to pass through them than denser bamers. The displaced

flow over the top is reduced on a more porous bamer. Similarly, the airfiow is disturbed

less at the point of the barrier, hence minimizing violent eddying, and creating a more

gentle flow of air at a reduced velocity. Konstyantinov and Stnizer (1969; cited in Bean et

al., 1975) argued that the amount of sheltering can also be influenced by the magnitude of

the wind velocity. Further, they stated that higher wind velocity increases the efficiency of

low porosity wind-barriers. and decreases that of the higher porosity, while

medium wind velocity has no beanng on the amount of sheltering.

The optimum porosity or permeability is of great importance in addressinç t

ight to

ie design of

the shelter. As noted above, a barrier with optimum porosity is suficiently permeable

throughout its height to allow air to pass through. This will result in a minimum of air

turbulence on the leeward side, and a maximum distance for windspeed reduction. In this

regard, however, there are conflicting opinions about the optimum porosity of barriers.

These could be attnbuted to the limitations in the study of porosity in early work . For

example, Hagen and Skidmore (1971) considered 35 to 45% to be an optimum porosity,

while Read (1964) concluded that 40 to 70% was optimum for field windbreaks.

Nevertheless, there seems to be a general acceptance that windbreaks with porosity of 40

to 50% will produce the largest zone of wind reduction, with minimal turbulence. This

range is optimal for crop and soil protection, but a single design does not meet al1

possible desired objectives. For example, for livestock shelter and snow management, a

lower porosity should be considered in order to obtain maximum eficiency (Frank et al.,

1 W6b; Peterson, 1984).

Windbreaks for snow management should be dense to evenly distribute the snow so that it

will contnbute to soi1 water recharge upon rnelting. to protect over-winter crops. A study

by Peterson (1984) on s h b bamers showed that a porosity of 33%, gave the most

uniform snow distribution. If this is the case, then overgrown natural fencerows with low

porosity could be of importance in snow management and livestock protection.

Effects of Shelter on Crop Yields

One of the most important uses of field windbreaks has been their role in protecting crops

from adverse winds. Indeed, the intensive use of shelter in Canada might have been

related to severe crop loss due to soil erosion in the Canadian Prairies (Rosenberg, 1983).

Likewise, to protect the decline of fencerows in Ontario farmers must be convinced to

change their misconception of fencerows as being of no economic value. They must be

shown that fencerows can provide an economic benefit through improved crop yield.

Thus, in this document, shelter effect on yield is included to show its potential impact on

crop productivity.

Wind has a pronounced effect on the growth and development of plants. In certain

locations, prevention of physical damage from frequent or occasionally high winds is the

most important feature of windbreaks. Mechanical damage to plants, such as distorted

shape and reduced sizes is observed. Shelter may indirectly infiuence crop yield through

its effects on insects and diseases floral initiation, pollination, and flower and fmit drop.

However, even when the above factors are not involved, there is a considerable evidence

in the literature that windbreaks alter microclimate, creating favorable conditions for crop

growth and yield (Rosenberg, 1966; Grace, 1977; Baldwin, 1988; Zhang, 1993).

Although these studies showed divergent amounts of increase in growth, early maturity,

higher quality, and greater econornic gain for various agricultural crops, the use of

extensive wind protective measures was justified.

Yields of vegetables are improved between windbreaks, compared to the open fields. For

example, in Australia a study to quanti@ the benefit of shelter to the production of potato

(Solatnim ~ziberostrnt) showed a 6.7% increase in yield behind a Ezicnlyprzrs shelterbelt

with a porosity of 44% (Sun and Dickinson, 1994). Monnette and Stewart (1987)

dernonstrated that windbreaks intluenced the vegetative development of pepper

(Capsiczmz a)mirm) increase in both fresh weight, and marketability of fmit.

In Quebec, an average yield increase of 27% was documented for tomato (Lycopersicort

esnrlen~zrm) grown behind a windbreak, as compared to the control fields (Fortin, 1986).

The windbreak increased crop quality, and the percentage of marketable fmits up to 30H

behind the windbreak. Similar studies on tomato, and snap beans (Phczseolzis vztlgaris)

demonstrated accelerated vegetative growth, earlier ripening of fniit, and a 16% to 44%

yield increase (Bagley, 1964).

Field Crops

A five year study in Saskatchewan to determine the effect ofwindbreaks on wheat

(Trifiam aes f im) yield demonstrated a 24% to 43% yield gain above the control

(Pelton, 1966). However, yields varied from year to year, suggesting the effect of other

environmental factors on growth. Similarly, Staple and Lehane (1955) found a net

increase in yield of 0.7 bushels, taking into consideration the area taken by the trees of the

windbreak.

The effect of windbreaks on soybean (Glycirze m m ) yield has been studied more than

other field crops (Radke and Burrows, 1970; Frank et al.. 1974a; Radke and Hagstrom

1974; Ogbuehi and Brandle, 198 1; Rasmussen and Shapiro, IWO). These studies are of

great importance as soybean is extensively grown in Ontario. Frank and coworkers

(1974a) study on soybean, under dryland and imgated regirnes, showed an increase of

20.4 - 24.0 hL/ha for imgated fields, and 1 1.8 hUlia - 12.8 hL/ha increase for dryland

agriculture, for sheltered plants over unsheltered controls. In addition, dry matter

production, green leaf area, and plant height were generally higher when soi1 water was

not a limiting factor. Similarly, Radke and Burrows (1 970) found soybeans sheltered by

temporary windbreaks grew taller, produced more dry weight, and higher grain yield

between 6H to 1 SH. Ogbuehi and Brandle (198 1) working on rainfed conditions, found

that yield around shelters was 20 to 26% higher than the exposed plots, suggesting that

increased water-use eficiency, and a significant increase in soybean yield can be expected

in shekered conditions.

Some workers have found no effiect on some crops grown under shehered conditions. For

example, two year results from Senthinathan and coworkers (1984) showed no beneficial

influence of windbreaks on the total water-use, rate of phenological development, height,

leaf area, top dry weight, yield cornponents, or seed yield of soybean. Greb and Black's

(1958) data showed a yield reduction in wheat and sorghum (Syricz~rrt gratnim). Although

they were not conclusive, they attnbuted the reduction in yields to water and nutrient

competition from the trees of the windbreaks. Indeed, a decrease in growth and yield

close to the windbreak is usually considered a result of competition. To avoid this

problem, Rasmussen and Shapiro (1990) suggested root pmning, while Iafri and

coworkers 0991) suggested improved cultural practices and control of weeds. The study

of Rasmussen and Shapiro (1990) on tree root-pmning adjacent to windbreaks showed

increased yields of up to 35% for corn and soybeans. Cutting of lateral tree roots

extending from a windbreak into the crop field has the potential to reduce competition for

moisture when it is a lirniting factor.

This study was undertaken to investigate the effect of natural fencerows on windspeed, air

temperature and relative humidity , and to study the porosity which are considered to be

an important factor in the eficiency of these fencerows

MATERIALS AND METHODS

The study was conducted on two sites in Eastern Ontario: Lunenberg in 1995, and

Metcalfe in 1996 (Figure 1). The general rnethods used were similar, at both sites. where

there are differences, this is mentioned.

The project was initiated in 1995, with the identification of the Lunenberg site, and the

collection of the windspeed, air temperature, and relative humidity information. The data

were later analyzed to observe the effect of the present fencerow on these microclimate

parameters. In the second year the investigation was broaden by collecting more field data

from another site, and using different instruments. Thus, in 1996, an additional site at

Metcalfe was identified, and the data collection process was repeated.

Site Selection

More than 1,000 aerial photographs taken 199 1 were examined at the Natural Resource

Information Centre in Toronto to identi@ potential project sites. Based on this

examination a number of candidate fencerows were visited. Factors considered in the

process of choosing a suitable study site included a relatively homogeneous fencerow with

respect to species composition, height, width, and the surface roughness of the adjacent

fields.

The first site selected was a natural fencerow with an east/west orientation, and was

identified as the first project site at Lunenberg, south of Ottawa (Figure 2, and Appendix

1). This fencerow consisted mainly of tree species such as basswood (Tilin anrerica~la

Figure 2. Aenal photo of the Lunenberg. The arrow indicate the study fencerow.

L.), trembling aspen (Popzilus tremzdoideS Mich.), common apple (Malm spp.) and ash

(Fraxbircs. spp.). The fencerow at Lunenberg nad a length of 184 m, with an average

height of 8 m, and an average width of 5 m. The second site selected for data collection

was at Metcalfe, Ontario (Figure 3, and Appendix 2). Fencerow length at this site 196 m,

with an average height of 10 m, and an average width of 6 m. The orientation and the

species composition of this fencerow was relatively similar to that of Lunenberg.

Data ColIection

To correct for inter-instrument differences, al1 6 meteorological stations were grouped

together, and data were collected for over 48 h. Ali rneteorological parameters

(windspeed, temperature, and relative humidity) were recorded at the same point for a

period of two days, to provide calibration data.

Subsequently, windspeed air temperature, and relative humidity readings were collected 1

m above the ground along transects running perpendicular to the fencerows. Average tree

height was determined using a hypsometer and stations established at multiples of height

(H). Six stations were set up at distances of 1H ( lx the average height of the windbreak),

5H, 10H, 15H, 20H, and 25H, along three transects perpendicular to the selected

fencerow. The furthest station, positioned at a distance of ZH, was assumed to be

beyond significant influence of the fencerow.

Data were collected at the Lunenberg site from July to August 1995. Windspeed and

direction were recorded continuously with an anemometer and a wind-vane (Lander

Figure 3 . Aerial photo of Metcalfe project. The arrow indicate the shidy fencerow.

Control System, Canada) at each of the stations, along three transects. Dry and wet bulb

temperature probes (Lander Control System Inc.) were utilized at each of the station to

measure air temperature and relative humidity. Temperature probes were housed in

Stevenson Screens secured on wooden posts. Meteorological data were recorded

continuously every 15 minutes, for two days (at Lunenberg), on each of three transects,

using CR10 data loggers (Campbell Scientific, Canada).

Likewise, from August to September 1996, the second set of data was collected at the

Metcalfe site. A similar procedure was used with the followinç exceptions. For this site,

207F (Campbell Scientific, Canada) temperature probes, which record air temperature and

relative humidity simultaneously, were utilized to monitor these parameters. The data

were again recorded every 15 minutes but for four days, on each of the five transects. Due

to logistic problems no modeling of days in terms of weather condition (windy, sunny,

rainy, etc.) were conducted. However, al1 data at each transect was continuousIy recorded

for two days at Lunenberg and four days for Metcalfe over the summer of 1995 and 1996,

respectively.

Porosity Estimation

Colour photographs of the various segments (plots) of the fencerows were taken at a

perpendicular distance of approximately 1H on a ciear day . These pictures were scanned

electronicdly, and edited to create a mosaic image of the each fencerow. Using the

software Micrografix Picture Publisher (1994) different shades and colours were

converted to black (foliage, branches, and tmnks) and white (pores). The images were

then converted to grids (gnd raster files) using ArcView GIS (Version 3.0a). The images

were randomly divided into segments or "plots" of equal length. The height of the plots

was equal to the average height of the fencerow. By dividing the fencerow into plots (9

for Lunenberg and 7 for Metcalfe), it was possible to estimate the variation in porosity

along the fencerow. Subsequently, using ArcView 3.0a each ce11 of the grid was assessed

to determine if it consisted of foliage trunks or branches (black), or fell on pore space

(white). An estimate of the porosity for each of the photo-plots was deterrnined using this

information.

Data Analysis

The data collected at the same point for a period of two days was used to correct for

inter-instrument differences. For this purpose, one instrument was selected as a reference,

and linear regression was used to develop correction terms for each of the remaining

instruments. Al1 subsequent data were adjusted by this correction factor using the

regression model:

~=Po+Pix

Where:

Y = the corrected value for windspeed, temperature, or relative humidity;

P o = the intercept;

P I = the dope; and

X = the uncorrected value

m e r the correction for inter-instrument differences, these data were surnmarized by five

readings bracketing each of six nominal hours, for each transect. The nominal hours were

0400h, 0800h, 1200h, 1600h, 2000h and 2400h. For example, the value for 0400h was

the mean of the values at 0330h, 0345 h, 0400h, 04 15h, and 0430h.

Al1 the corrected meteorological data for both sites were expressed in relation to the data

for the farthest station (25H), which is assumed to be beyond the influence of the

fencerow. This approach to expressing the absolute values in terms of relative

(standardized) values reduces the impact of other factors that are beyond the influence of

the fencerow. These data are referred to here as: "standardized windspeed" (SWS),

standardized temperature (ST), and standardized relative humidity (SRH). To achieve

this, the values at each station were divided by the corresponding value of the farthest

station (25H). For instance, for the standardized value of windspeed (SWS) at 1200h for

station 1, the absolute value at this and tirne was divided by the absolute value at station 6

at 1200.

Although the method applied at each site is essentially the sarne, the physical structure of

the fencerows, differences in local environment, and differences in instmments, made it

inappropriate to make statistical cornparisons between the two sites.

Al1 meteorological data (standardized windspeed, air temperature, and relative humidity)

were analyzed using analysis of variance (ANOVA) procedures (GLM of SAS), with

"transects" as the replicates (Table 1).

The following linear model was used in the ANOVA:

Where:

Yijk(jL(I)= the response of the experimental unit on day (i), hour (j) at position (k) for replicate

(1) (transects);

p= Overall mean;

Di= effect of day i (i = 1 and 2) for Lunenberg and (i = 1 .... 4) at Metcalfe;

Hj= effect of hour j (j=l . . . 6 )

&= the effect of station (position) (k = 1.. . . -5);

DHij= the effect due to the interaction between days(i) and hoursa);

DSij = the effect of the interaction of the days (i), and station (k);

SHjk= the ef5ect of the interaction of the station (k), and hour (j);

DHSijk= the effect of the interaction of days, hours and stations;

Eijii(l)= the experimental error associated with experirnental unit (1) on day (i), hour (j), and station@)

The expected rnean squares (EMS) for this linear model are listed in Table 1.

Table 1 . Expected mean square for each source for standardized windspeed, temperature and relative hurnidity used to test the hypothesis. Source Degree of Freedom @F) Expected means square (EMS)

Di (i- 1) 0.2 + pado Hj (i-1) a.Z +6y~20H + a6ya2H S j (k- 1

2 OC +By&S + CLPYQ~S

DHij (i- 1)G- 1) CE + 6yG2~ii

DSik (i- l)(k- 1) d ,'+ PY&S Asjk (j- l)(k- L) o ~ + a y o 2 ~ DHS,~ (i- 1)~- 1)k- 1) oCw20r ls

($cl)- 1 2 E GO.) GE

When statistical significance was identified by the ANOVq Duncan's Multiple Range Test

(a=0.05) was used to separate the means. When applicable, the results of these tests are

shown in the figures. Before the ANOVq normality was tested using the W-test.

Homogeneity of variance was also tested using Bartlett' s test.

RESULTS

A number of the main effects used in the ANOVAs are not discussed. They were

included in the analyses to explain more of the variation (and reduce the residual error)

and had no bearing on the results at hand.

The main effects hour, days, stations, and their interactions effects are considered for the

Andysis of Variance model for the parameters SWS, ST, and SRH and are included in the

Appendices. However, one of the main effects hour used in the model is generally

expected to exhibit significant differences because of the variation of these microclimatic

parameters during different times of the day. Thus, "houf' is not discussed in this

document. Only day , station, and the interaction day x station are considered in the

discussion.

The average monthly meteorological data fiom the nearest Environment Canada weather

station (üpland's Airport, Ottawa) for the months of July, August, and September of 1995

and 1996 are surnrnarized in Appendix 3.

Standardized Windspeed (SWS)

Figure 4 shows the standardized windspeed pattern for the Lunenberg site. The lowest

windspeed was recorded at the first station (lH), and the highest SWS was recorded at

the last station (20H fiom the fencerow).

No significant differences were found arnong days (P = 0.9542), or stations (P = 0.3491).

The interaction between these two main effects was also not significant (P = 0.5804). The

results of the analysis of variance for SWS at Lunenberg are summarized in Appendix 4.

Figure 5 represents the standardized windspeed pattern at the five stations for Metcalfe

site. The lowest mean standardized windspeed was at IOH, and the maximum at 1H and

20H. Analysis of variance revealed a significant difference (P = 0.0177) among the

stations, indicating that the SWS was afEected by distance fiom the fencerow (Appendix

5). The variation due to day (P= 0.1455), and the interactions between the day and

stations (P= 0.9922) were not significantly different.

Standardized Temperature (ST)

The lowest temperature at Lunenberg (Figure 6) was recorded at the station fûrthest fiom

the fencerow, at 2OH. The result of the analysis of variance for standardized temperature

(ST) showed a significant variation arnong the stations (P = 0.000 l), and the interactions

between the station and day (P = 0.0012) (Appendix 6 and 7). No significant differences

was detected between day (P = 0.8696).

At the Metcalfe site (Figure 7) the highest ST was recorded at the first station (lm and

Figure 4.

Figure 5.

Standardized wuidspeed (SWS) profile as affected by the distance (stations) from the natural fencerow at Lunenberg.

2.00

A AB B AB A 0.00 ! 1 T L 1

1H 5H 1 OH 15H 20H 25H

Distance (stations)

1.50 -

cf3 1.00 -

Standardized windspeed (SWS) profile as affected by the distance (stations) from the natural fencerow at Metcalfe.

.-.---- ---.*. C

0.50

0.00 Y I 1 I 1 I I

I 1

1H 5H 1 OH 15H 20H 25H

Distance (station)

Distance (station)

Figure 6. Standardized temperature (ST) as affected by the distance (stations) fiom the natural fencerow at Lunenberg.

Figure 7.

Distance (stations)

Standardized temperature (ST) as afEected by the distance (stations) fkom the naturai fencerow at Metcalfe site.

the dependent variable for ST indicated a significant difference (P = 0.0025) arnong the

stations (Appendix 8). No variation was found between day (P= 0.7408), or for the

interaction between the day and station (P= 0.9885).

Standardized Relative Humidity (SREI)

The average SRH for two days for the Lunenberg site are illustrated in Figure 8 . No

significant differences were found between days and among stations. The interaction

between the two also found to be non-significant (Appendix 9). This indicates that

distance (stations) did not affiect the SRH for the Lunenberg fencerow.

Figure 9 shows the standardized relative humidity for the Metcalfe site. At Metclafe

S R H was significantly different among stations (P = 0.0001) (Appendix 10). There was

also a significant difference among days (P = 0.00 19), but the interaction between days

and stations was not significantly diKerent (P = 0.78 15). The SRH was affected by

distance.as the highest SRH recording was at station 2 (lOH), and the lowest at station 3

(10H). The SRH recorded at these stations are significantly different fiom the stations 1

and 4.

Figure 8. Standardized relative hurnidity (SRH) as affected by the distance (stations) from the naturai fencerow at Lunenberg.

2.00

1.50 -

m

0.50 -

0.00

Distance (stations)

--...---- w - v

1 r 1 1 1

Figure 9.

1H SH 1OH 1SH 20H 25H

Distance (stations)

Standardized relative humidity (SRH) as affected by the distance (stations) fiom the natural fencerow at Metcalfe site.

Porosity Estimation

Figure 10 iiiustrates the approximate position of the sarnple plots used in the estimation of

porosity for the Lunenberg site. Porosity ranged f'iom 13 to 54%, with a mean of 30 %

and a coefficient of variation (CV.) of 0.45 (Table 2).

Table 2. Estimated porosity of nine plots dong the fencerow at the Lunenberg site.

PLOT # 1 2 3 4 5 6 7 8 9 ~ A N C V POROSITY(%) 21 22 22 40 45 22 54 13 30 30 0.45

The approximate positions of the seven plots used to estimate the porosity of the Metcalfe

sites are shown in Figure 11. The results of the porosity calculations at the Metcalfe site

showed a range in porosity of 24% to 46%, with a mean of 33%. Table 3 shows the

values for percentage porosity for the seven sarnple plots.

Table 3 . Estimated porosity of seven plots dong the fencerow at the Metcalfe site PLOT # 1 2 3 4 5 6 7 M'AN C.V POROSITY (%) 46 37 29 36 24 25 32 33 0.23

The coefficients of variation for the fencerows' porosity express the degree of variation

among the photo-plots. For example, the values of the coefficients of variation (which are

0.45 and 0.23 for Lunenberg and Metcalfe, respectively) signifi a higher variation in the

porosity of the fencerow at the Lunenberg site.

A t-test was used to test the hypothesis that there is no significant difference between the

mean porosity of the two fencerows. The hypothesis could not be rejected (P = 0.607)

suggesting that the recorded diEerences are not significant.

Porosity % 2 1 22 22 40 45 22 54 13 30 Plot 1 2 3 4 5 6 7 8 9

Figure 10. Digitized photo of the Lunenberg fencerow used in the estimation of nine sample plots along the fencerow.

Porosity % 46 Plot 1

Figure 1 1. Digitized photo ofthe Metcalfe fecerow used in the estimation of seven sample plots along the fencerow

DISCUSSION AND CONCLUSION

In two separate field studies (Lunenberg and Metcalfe) on the effect of n a ~ r a l fencerow

on the microclimate factors it has been demonstrated that the SWS was not significantly

different for Lunenberg, and that SWS was significantly altered for MetcaEe. In addition,

in the study of the physicd characteristics of these fencerows porosity variation was

observed.

In a theoreticai framework for studying field shelter, it is important to note that it is

extremely complicated to fully understand the physics of wind and the accompanying

microclimate effects. Furthemore, the relatively limited published information on

microclimate is confined to planted windbreaks in fields, and windtunnel experiments.

Microclimate Effect

Theoretically windspeed adjacent to an efficient windbreak tends to increase gradually

with increasing distance h m a bamer (Grace, 1977; McNaughton, 1988). To a certain

degree the result fkom both sites in the current study showed a pattern supponing this

generd theory. However, the results were not always consistent, particularly at

Lunenberg where there was no signifiant difference among the stations in SWS. The

SWS pattem showed a reduction at stations 1H and 5H, and an increase at 10H before

dropping at 15H. The windspeed increased further downwind at 20H (Figure 4). This

shows some pattem of reduction of windspeed at the nearest station, and augmentation at

the last station, but the results ffom the intermediate stations did not follow the expected

trend.

The fencerow at the Metcalfe site exhibited significant differences among the stations (P=

0.0 177)with respect to SWS. The hypothesis that there was no significant difference

among stations with respect to SWS must be rejected. The lowest SWS was found at

station 3 (10H) suggested a significant reduction of SWS at this point. The mean values

for stations 2 (SH), 3 (lOH), and 4 (15H) were significantly different from the fifth station

(20H), indicating a significant reduction of SWS at these points before the SWS increased

fiirther at 20H. However, it is important to note that the SWS at the first station (1H) was

not significantly different fiom the last station, showing that a reduction in SWS did not

commence irnmediately f i e r the fencerow (1H). This rnight be due to a possible

turbulence effect a d o r rnight be associated with the physicai structure of the fencerow.

The standardized temperature (ST) for the Lunenberg site showed a significant difference

arnong stations (Table 5). The highest ST values were recorded at station 3 and 4 (Figure

6). Sirnilarly, the standardized temperature at the Metcaife site also showed a significant

difference stations (Table 6), with the highest average ST at station one, and the lowest

value at the farthest station, at 20H (Figure 7). This trend is consistent with earlier work

that suggested a higher daytime and Nghttime temperature closer to the shelter (WoodniR

et al., 1959; Read 1964).

The SRH for Lunenberg site showed no significant difference arnong the stations.

Taking into consideration the nonsignificant difference of relative windspeed at

Lunenberg, the SRH effect would not be expected to be strong. In contrast, to the

Lunenberg fencerow, the SRH at the Metcaüe site showed a significant difference, with

the highest mean SRH at station 2 (10H) close to the fencerow, and the Iowest mean at

the fifth station, positioned at a distance of 20H. This result not ody indicated a

significant effect of distance (stations) on humidity, but also that SRH was higher closer to

the fencerow, and lower further downwind.

The effect of a fencerow on temperature, or relative humidity is not expected to be

significant if there is no eEect of the shelter on windspeed. At Lunenberg, the shelter

effect on windspeed was not significant. Nevertheless, the low windspeed recorded at the

first station, and highest speed at the last station followed the expected trend.

Furthermore, the impact of shelter on SWS, ST, and SRH at the Metcalfe site supports the

idea that natural fencerows can be used as windbreaks.

Physical Structure (Porosity)

The ultimate goal of research on shelter is to delineate the most effective stmcture for

windbreaks. The design or the structural quality of shelter has a major impact on the

ability of the shelter to achieve the desired objective(s). The structural quality is a

function of the length, height, width, orientation, and, most importantly, the porosity of a

windbreak (Cabom, 1965; Rosenberg, 1966; K e ~ e y , 1986). For most purposes, a

windbreak or fencerow should not block the prevailing wind completely, but rather filter

the wind so that it is not forced to go over the top of the windbreak, or around its ends.

By redirecting the wind, rather than slowing it, turbulence may be created and the shelter

extent and quality will be reduced. This phenornenon makes the porosity of windbreaks,

or fencerows, an important function in relation to their efficiency.

For the Lunenberg fencerow, the porosity ranged between 13% and 54%, with a mean and

coefficient of variation of 3 0% and 0.45, respectively. Sirnilarly, the Metcalfe fencerow

had a range in porosity between 24% and 46%, with a mean and coefficient of variation.

of 33% and 0.23 respectively. Although, the optimum porosity of shelter has been

controversial, it is generally accepted that the maximum extent of shelter is when there is a

porosity of 40% to 50% (Cabom, 1965, Hagen and Skidmore, 1971; K e ~ e y , 1986).

Therefore, the rnean porosity for the two fencerows were slightly lower than optimum,

and not significantly different from each other. However, the respective CVs indicated

that the Lunenberg fencerow was much more heterogeneous

Figures 10 and 1 1 (the digitized mosaic pictures) illustrate the Lunenberg and Metcalfe

fencerows. Porosity varied from a relatively dense, to a more porous condition along the

fencerow. this was true in spite of the fact that areas of higher porosity (areas not

representative) of the study fencerows were avoided when the instmments were moved

dong randornly selected transects. The irregularity of porosity along the fencerows might

have affected the wind profile. Furthexmore, research has s h o w that under varying

porosity conditions, different wind profiles have been reported. For example, a

moderately porous windbreak, showed an effective reduction of windspeed to a distance

of 15H to 20H (Kemey, 1986). Heavily foliated trees or denser bamers exerted greater

drag and absorbed more momentum from the mean wind flow thus, creating turbulence

(McNaughton, 1988). The combined effect of these two scenanos dong the length of the

study fencerow may account for the failure to observe the anticipated patterns at stations

adjacent to the Lunenberg fencerow. The somewhat stronger trends noted for the

Metcalfe site, might be a reflection of somewhat less variation in porosity. However, it is

difficult to make any conclusive statement with respect to this, since it was necessary to

use different weather instruments between sites (years). In spite of my efforts to calibrate

the equipment, the greater variation in microclimate data at the Lunenberg site may have

been attributable to less precise instrumentation. Furthemore, the lirnited number of days

of the data collection might have contributed to the variation.

Conclusion

The best approach to investigate the effect of shelter on microclimates is to measure the

panuneters (air temperature, and relative humidity) which are directly affected by the

windspeed at different points from the shelter. The present study undertook this

approach, and attempted to assess the effect of a fencerow on the nearby microclimate.

Broadly speaking, the data showed trends in wind-patterns adjacent to the fencerows. At

both sites, the trends in SWS agreed with other studies that showed windspeed decreases

at certain distance, and eventually recovers as distance from the barrier increases

(Mulheam and Bradley, 1976; Grace, 1977; McNaughton, 1988). Similady, the

relationships between distance and the other two microclimate parameters ST and SRH

showed trends similar to those exhibited by planted windbreaks, and artificial barriers.

However, at the Lunenberg site these trends were not strong, which may be explained by

variability in the porosity of the fencerow.

As show here, natural fencerows probably exhibit more variation in porosity than would

be expected in most planted windbreaks. If üiis were the case, then the variation in the

structure of fencerows could be seen as a drawback wher. considering them for shelter.

On the other hand, this could be considered an asset as a wildlife habitat, comdor

establishment, or aesthetic purposes (the latter may be a matter of personal preference).

If the efficiency of windbreaks is afliected by their physical structure, and in particular their

porosity @oth in absolute tenns and in variability), then it seems reasonable to assume that

some management regirnes could be applied to upgrade the general structure of the

fencerows. This was beyond the scope of the study, to suggest how to apply any such

regirne, however it would have been very interesting to monitor the wind profile after

conducting some management operations.

Findly, as mentioned in the literature review, various shelter and microclimate studies

have been conducted over the years, in different parts of the world. The approaches used

in these studies differed considerably. Thus, the results of this kind of study have often

differed, leading to variable and sometimes conflicting results. Apart from the cornplexity

of the mechanism of shelter, the variations in reports can also be attributed to the very

physical nature of the shelter structures being studied. This could be further complicated

by the local environment in which the study was conducted, the air flow regime during the

study penod, the location and type of equipment used, and the interpretation of the result.

RESEARCH NEEDS

Some of the areas of fiiture research that should be undertaken in order to enrich the

knowledge of the effects of fencerows on fencerows on microclimates are listed below.

To conduct studies over longer penod of time (season) incorporating the design

and vegetation composition of the fencerow.

To develop guidelines for different management regimes to improve the efficiency

of natural fencenvos in the province.

To develop recommendations for selcction and establishment of tree species to

modify the natural fencerow physicai structure.

To investigate the effect of fencerows on plant growth, in order to convince

f m e r s to keep fencerwos in their farmiands.

To investigate natural fencerow microclimate studies with its importance to

biodiversity, wildlife habitat, and other important uses.

LITERATURE CITED

Aase, J.K., and F.H. Siddoway. 1974. Tai1 wheatgrass barriers and winter wheat response. Agric. Meteorol. l3 :X 1-33 8

Bagley, W.T. 1964. Response of tomatoes and beans to windbreak shelter.

J. Soil Water Conserv. News l9:7 1-73.

Baldwin, L.S. 1988. The influence of field windbreaks on vegetable and speciality crops. Agric. Ecosystems Environ. 22/23: 19 1-203.

Bates, C.G. 191 1. Windbreaks: Their influence and value. USDA For. Serv. Bull. No. 86-100 pp.

Bates, C.G. 1944. The windbreak as a farm asset. USDA F m . Bull. No. 1405.16 pp.

Bean, A., Alperi, R. W., and Federer, C.A. 1975. A rnethod for categorizing shelterbelt porosity. Agr. Meteorol. 14:4 17-429.

Bennett, A.F., Henein, K., and Memam, G. 1994. Conidor use and the elements of comdor quality: Chipmunks and fencerows in a farmland mosaic. Bio Conser. 68: 155-165.

Berna, F.D. undated. Managing existing fencerows for windbreaks. USDA Soil conser. Service, Indiana.

Best, B.L. 1983. Bird use of fencerows: Implications of contemporary fencerow management practices. Wildl. Soc. Bull. 11(4):343-347.

Borrelli, J., G., J.M., and Abtew, W. 1989. Wind bamers: a revaluation of height, spacing, and porosity. Trans. ASAE 32:2023-2027.

Brown, K. W., and Rosenberg, N.J. 1970. Shelter eEect on microclimate, growth and water use by imgated sugar beets in the Great Plains. Agri Meteo. 9:225-240.

Cabom, J.M. 1957. Shelterbelts and microclimate. U.K. Forestry Comm. Bull. No. 29. 200 pp.

Cabom, J.M. 1965. Shelterbelt and windbreaks. Faber and Faber Ltd. London: 288 pp.

Dewalle, D.R., and Heisler, G.M. 1988. Use o f windbreaks for energy conservation. Agric. Ecosystems Environ. 22/23:243-260.

Fortin, P.A. 1986. Windbreak effect on growth and yield of tomato in Quebec. M.Sc. Thesis.

Frank, A.B., Harris DG., and Willis, W.O. 1974 a. Windbreak influence on water relations, growth, and yield of soybeans. Crop Scie. 14:761-765.

Frank, A.B., Hams D.G., and Wi!Iis, W.O. 1974 b. Influence of windbreaks on crop performance and snow management in Northem Dakota. In Proceeding of the Symposium, Denver, Col., Apd 20-22, 1976. Great Plains Agric. Counc. Pub. No.78. 218pp

Gloyne, R.W. 1954. Some effects of shelterbelts upon local and microclimate. Forestry, 27:85-95.

Grace, JI 1977. Plant responses to wind. Academic Press Inc. London.

Greb, B.W., and Black, AL. 1958. Effect of windbreak plantings on adjacent crops. ~~1223-227.

Guyot, G. 1963. Windbreaks: Microclimate modification and arnelioration of agncultural production. Am. Agron. 14:429-488.

Hagen, L.J., and Skidmore, E.L. 197 1. Turbulent velocity fluctuations and vertical flow as affected by windbreak porosity. Trans the ASAE 14(4):634-637.

Jensen, M. 1954. Shelter effects. Investigation into the aerodynamics of shelter and its effect on climate and crops. Danish Tech. Press. Copenhagen. 266pp.

Johnson, R.T., and Beck, M.M. 1988. Influence of shelterbelts on wildlife management and biology. Agric. Ecosystems Environ. 22/23: 3 0 1 -3 3 5.

Kemey, W. A. 1 987. A method for estimating windbreak porosity using digitised photographie silhouettes. Agric For Meteor01 39:9 1-94.

Kemey, W.A. 1986. Studies on the design and management of shelterbelts and windbreaks in Southem Ontario. M.Sc., Thesis.

Konstantinov, A.R., and Stnuer, L.R. 1965. Shelterbeits and crop yieids. USDA and Nat.Sc. Found. Israle Prog. For Sc. Trans.

Loeffler, A.E., Gordcn, A.M. and Gillespie, T.J. 1992. Optical porosity and windspeed reduction by coniferous windbreaks ir. Southem Ontario. A.groforestry System, 17: 119-133.

McNaughton, 1988. Effect of windbreaks on turbulent transport and microclimate. Agric.Eco-systems Environ., 22/23: 17-39.

Miller, D.R., Bagley, W.T., and Rosenberg, N.J. 1974. Microclimate modification with shelterbelts. J. Soil Water Cons. -- 41-44.

Monnette, S., and Stewart, K.A. 1987. The effect of windbreaks and mulch on the growth and yield of pepper (Capsicm artmrunz f) . Cm. J. Plant Sci., 67:3 15-320.

Ogbuehi, S.N., and Bradle, J.R. 198 1. Idluence of windbreaks-shelter on soybeans production under rain-fed conditions. Agron. J. 73:625-628.

Mulheam, P.J., and Bradley, E.F. 1977. Secondary flows in the lee of porous shelterbelts. Boundary-layer Meteorol. 12:75-92.

Pelton, W.L., 1976. Windbreak on studies the Canadian Prairies. Proc. Syrnp. Shelterbelts on the Great Plains. Denver, Colorado. Great Plains Agric. Concil Pub. N0.78. pp. 64-68.

Peterson, T.C. 1984. Outdoor scale modelling of s h b barriers in drifting snow. Agric For Meteorol. 3 1 : 167- 18 1.

Radke, J.K., and W.C. Burrows. 1970. Soybean plant response to temporary field windbreaks. Agron. J. 62:424-429

Radke, J.K., and Hagstrorn R.T. 1974. Wind turbulence in a soybean field sheltered by four types of wind bamers. Agro. J. 66: 273-278.

Rasmussen, D.S., and Shapiro, A.C. 1990. Effect of tree root-pmning adjacent to windbreaks on corn and soybeans. J. Soi1 and Water Cons. 45571-575.

Read, R.A. 1964. Tree windbreaks for the Centrai Great Plains. US. Department of Agriculture Handbook. 58-67 pp.

Rosenberg, J. N., 1966. Microclimate, air mWng and physiological regdation of transpiration as influenced by wind shelter in an imgated bean field, Agr. Meteorol. 3: 197-224.

Rosenberg, J.N. 1983. Microclimate: The biological environment. Wiley- Interscience Publication, United States. pp. 33 1-368..

Senthinathan, A., McVetty, P.B.E., and Stobbe, E. H. 1984. Soybean plant response to amuai crop windbreaks under rain-fed conditions. Can. I. Plant Sci. 64: 879-883.

Shaw, D.L. 1988. The design and use of living snow fences in North Amenca. Agric. Ecosystems Environ. 22/23:3 5 1-3 62.

Spanbauer, M.K. 1989. Fencerows. -------.

Staple, W. J., and Lehane, J.J. 1955. The influence of field shelterbelts on wind velocity, evaporation, soi1 moisture, and crop yield. Can. J. Agr. Sci. 35: 440453.

Sun, D., and Dickinson, G.R. 1994. A case study of shelterbelt effect on potato (Solunum

tuberosum) yield on the Atherton tablelands in tropical north Australia. Agroforestry-system. 25: 2, 14 1-1 5 1.

Van der Linde, RJ. 1962. Trees outside the forest. In Forest Influences. FA0 Forestry Prod. Study No 15. pp 141-208.

Van Eimern., J.R., Karschon, J.R., Rmmova, L.A., and Robetson, G. W. 1964. Wmdbreaks and shelterbelts. WMO Tech. Rep. No.59. 200 pp.

Wegner, J.F., and Memam, G. 1979. Movements by birds and smdl marnrnals between a wood and adjoining f m land habits. J. Appl. Eco. 16:349-3 56.

Woodmffe, N.P., Read, R.A., and Chepil, W. S. 1959. Influence of a field windbreak on summer wind movement and air temperature. Agri Exp Station, Kansas State Uni. Tech. Bull. 100:2-24.

Zhang, Hehui. 1993. Windbreak shelter and physioiogical responses of corn. Ph-D. Thesis.

APPENDICES

Appendix 1 . Photographs of Lunenebrg fencerow

Appendix 2. Photograp hs of Metcaife fencerow

.- - --

Appendk 3. The Environment Canada monthly rneteorological summary of Ottawa Airport for the months of July, August and September for 1995 and 1996

(Source:Environrnent Canada).

Months Windspeed Air Temperature Relative Humidity Rainfall

ml ec) PW (mm)

h . 1 ~ Min Max Min Normal Total

July 1995

July 1996 10.7 25.4 14.8 84 49 88.1 88.1

Appendk 4. A summary of the ANOVA used to test the hypothesis (Ho) of no significant difference arnong days, hour, stations and the their interactions with respect to

standardized windspeed (SWS) for Lunenberg Fencerow Site.

Source DF Mean Square F Value P r > F

Days 1 0.0 1662722 0.00 0.9542

Hour 5 18.16201922 3.62 0,0044

Days * Hour 5 2.77693656 0.55 0.7357

Stat 4 5.63742417 1-12 0.3491

Days*Stat 4 3.6 1 168694 0.72 0.5804

Stat*Hour 20 2.9045 1950 0.58 0.92 1 O

Days* Stat*Hour 20 1.62209628 0.32 0.9974

Residual 120 5.02105500

Total 179 39.75236489

Appendix 5. A surnmary of the ANOVA used to test the hypothesis (Ho) of no significant difference among days, hour, stations and the their interactions with respect to

standardized windspeed (SWS) for Metcalfe Fencerow Site.

Source DF Mean Sqaure F Value Pr>F

Days 3 0.24407528 1.80 0.1455

Hour 5 8.20636337 60.67 0.000 1

Days * Hour 15 0.37127608 2.74 0.0004

Stat 4 0.40841917 3 .O2 0.0 177

DayseStat 12 0.03783639 0.28 0.9922

S tat*Hour 20 0.0269943 7 0.20 0.9999

Days*Stat*Hour 60 0.01 548586 0.11 1 .O000

Residual 480 O. 13526492

Total 599 9.44571544

Appendix 6. A summary of the ANOVA used to test the hypothesis (Ho) of no significant diierence arnong days, hour, stations and the their interactions with respect to

standardized temperature (ST) for Lunenberg Fencerow Site.

Source DF Mean Sqaure F Value P r > F

Days 1 0.00001389 0.03 0.8696

Hour 5 0.00028722 0.56 0.7308

Days * Hour 5 0.00026056 0.5 1 0.7701

S tat 4 0.00720472 1 4.04 0.000 1

Days* Stat 4 0.002479 17 4.83 0.0012

S tat*Hour 20 0.00052 139 1 .O2 0.4495

Days*Stat*Hour 20 0.00027250 0.53 0.9483

Residual 120 0.06 160000

Total 179 0.01 155278

Appendix 7. Standardized temperature (ST) as affected by the interaction of station and day for the natural fencerow at Lunenberg Site.

l H SH 10H 1SH 2 0 8 25H Distance (stations)

2.00

1 5 0 -

r+Day 1

+ Day 2

1

Appendix 8. A summary of the ANOVA used to test the hypothesis (Ho) of no significant dserence among days, hour, stations and the their interactions with respect to

standardized Temperature (ST) for Metcalfe Fencerow Site.

Source DF Mean Sqaure F Value Pr > F

Days 3 0.07947044 0.42 0.7408

Hour 5 0.48247787 2.53 0.028 1

Days * Hour 15 0.23 185324 1.22 0.2546

S tat 4 0 -793 76442 4.17 0.0025

Days*Stat 12 0.0580803 1 0.30 0.9885

Stat*Hour 20 0.3 1746962 1.67 0.03 53

Days*Stat*Hour 60 0.13902844 O. 73 0.9345

Residual 480 O. 19053583

Total 599 2.292680 17

Appendu 9. A surnmary of the ANûVA used to test the hypothesis (Ho) o f no significant daerence arnong days, hour, stations and the their interactions with respect to

standardized relative humidity (SRH) for Lunenberg Fencerow Site.

Source DF Mean Square F Value P r W

Days 1 0.02426722 1.32 0.2525

Hour 5 0.00563389 0.3 1 0.9079

Days * Hour 5 0.008 12456 0.44 0.8 179

Stat 4 0.02608806 1.42 0.23 10

Days* S tat 4 0.00 178250 0.10 0.9832

StatSHour 20 0.00149639 0.08 1.0000

Days*Stat4Hour 20 0.00137817 0.08 1.0000

Residual 120 0.01835389

Total 179 0.087 12468

Appendix 10. A surnmary of the ANOVA used to test the hypothesis (Ho) of no sigtilficant diKerence arnong days, hour, stations and the their interactions with respect to

standardized relative hurnidity (SRH) for Metcalfe Fencerow Site. - -- - - - - -

Source DF Mean Square F Value P r > F

Days 3 0.02421750 5 .O5 0.0019

Hour 5 0.02830217 5 -90 0.000 1

Days * Hour 1 5 0.00730537 1.52 O. 0924

Stat 4 0.73710067 153.71 0.000 1

Days*Stat 12 0.00320889 0.67 0.78 15

S tatSHour 20 0.03280617 6.84 0.000 1

Days* Stat*Hour 60 0.00 134626 0.23 1 .O000

Residual 480 0.00479525

Total 599 0.83908228

IMAGE EVALUATION TEST TARGET (QA-3)

APPLlED - I M G E . lnc = 1653 East Main Street - -- , , Rochester. NY 14609 USA -- ,==, Phone: 716/482-0300 -- -- - - Fa: 7 161288-5989

O 1993, Appiied Image, Inc.. Al1 Rights Reseiued