ESTIMATING WILDLIFE HABITAT TRENDS ON AGRICULTURALECOSYSTEMS IN THE UNITED STATES

Stephen J. Brady1 and Curtis H. Flather2

-- Plenary Session 2 --

Ecosystem/Habitats Impacted by Agricultural Activities

Tuesday 6 November 2001

Paper presented to the:

OECD Expert Meeting on Agri-Biodiversity Indicators5-8 November 2001Zürich, Switzerland

1 US Department of Agriculture, Natural Resources Conservation Service. Natural Resources Research Center, 2150Centre Ave.,Ft. Collins, Colorado 80526-1891, United States.

2 US Department of Agriculture, Forest Service Natural Resources Research Center, 2150 Centre Ave., Ft. Collins,Colorado 80526-1891, United States.

1

Estimating Wildlife Habitat Trends on Agricultural Ecosystemsin the United States

Stephen J. Brady1 and Curtis H. Flather 2

1US Department of Agriculture, Natural Resources Conservation Service2US Department of Agriculture, Forest Service

1&2 Natural Resources Research Center, 2150 Centre Ave.,Ft. Collins, Colorado 80526-1891

AbstractRecent trends (1992 – 1997) in wildlife habitat on agricultural and grazed ecosystems are reviewed usingdata from the 1997 National Resources Inventory. Land use changes, the losses and gains of wetlands, andreasons for wetland losses are described. Ecological indices describing spatial pattern and fragmentationof cropland and rangeland habitats are discussed, and geographically explicit summary statistics arepresented for the United States. The effect of the Department of Agriculture’s Conservation ReserveProgram is described as an example of a multi-purpose habitat intervention scheme. Because land resourceplanners need estimates of habitat quantity and condition, the concept of wildlife habitat management as asecondary use of agricultural lands is reviewed. The uses and limitations of ecological indicators andhabitat matrices, including statistical estimates of precision and the need to establish relationships betweenhabitat-based indicators and direct measures of biodiversity, are also discussed.

Keywords: Agri-environmental indicators, National Resources Inventory, wildlife habitat, United States.

IntroductionThe kind, amount, and distribution of life supporting elements on the landscape collectively determine thequality of habitat useful to wildlife. Each species has unique requirements with regard to these variouselements. Our knowledge of these requirements for most species is very limited while for a few speciesour understanding of their habitat needs is better. Species response to management of habitats exhibitsgreat variability over scales of both time and space. While our ability to measure ecological patterncontinues to improve there is much to be learned about ecological processes and functions. Althoughscientists continue to make models of ecological systems there remains a substantial amount ofunexplained variance in the explanatory power of those models. Habitat management is therefore acombination of both science and art. Aldo Leopold (1966: p. 177-178) provided an interesting description:“There is much confusion between land and country. Land is the place where corn, gullies, and mortgagesgrow. Country is the personality of the land, the collective harmony of its soil, life, and weather. ….Poorland may be rich country and vice versa. Only economists mistake physical opulence for riches. Countrymay be rich despite a conspicuous poverty of physical endowment, and its quality may not be obvious atfirst glance, nor at all times…. It [wildlife] often represents the difference between rich country and mereland.” All land can be considered habitat, but its quality varies from extreme poverty to abundance.

The task at hand is a challenging one, again as per Leopold (1933:p. 124): “When the game manager askshimself whether a give piece of land is suitable for a given species of game, he must realize that he isasking no simple question, but rather he is facing one of the great enigmas of animate nature. An answergood enough for practical purposes is usually easy to get by the simple process of noting whether thespecies is there and ready, or whether it occurs as ‘similar’ range nearby. But let him not be cocksureabout what is ‘similar’, for this involves the deeper questions of why a species occurs in one place and not

2

in another, which is probably the same as why it persists at all. No living man can answer that questionfully in even one single instance.”

Land use is the principal factor determining the base level of abundance of indigenous species. In mostcases land use has a greater impact on species abundance than does the management of land (e.g.,application of practices for agricultural production or soil and water conservation). Agriculture affectshabitat directly through converting natural habitats to cultivation, grazing, or other manipulation and theassociated repeated disturbances that accompany those conversions. Agriculture indirectly affects wildlifehabitat through water management practices for irrigation and drainage, soil erosion and sedimentation,and elevated nutrient and pollutant discharges into the environment. The direct effects of land useconversions on habitat are more easily measured than are the indirect effects. While the use of land isrelatively easy to document, assessing its quality (productive, economic, habitat, etc.) is more challenging.

Most indicators of agriculture’s affect on habitat reflect habitat patterns across the landscape. Thosepatterns and the biological diversity associated with them are the cumulative result of many ecologicalprocesses operating over time. It is much easier to describe the resulting patterns than it is to quantify theprocesses. Natural systems are inherently variable, and this variability is expressed both spatially andtemporally. Because wildlife populations are the result of many processes operating together, the quantityand quality of habitat are just two of the indicators affecting the distribution and abundance of wildlife.Population density is often an inaccurate estimator of habitat quality (Van Horne 1983) and somepopulation fluctuations are not related to habitat but may be the result of catastrophic weather conditions,disease, or overexploitation (Schamberger 1988). Habitat indicators may be useful for comparing betweenalternative agricultural land management scenarios for regional or national program planning and relatedpurposes. However, they must be designed so that they may be tested against empirically derived wildlifepopulation estimates. Ideally they will be designed as falsifiable hypotheses.

Testing of habitat indicators should be done with multiple measures of biological diversity to properlyreflect the complexity of natural systems. This is a problem because extant data sources representing manyspecies are usually not available. Where data are available for a class of organisms, such as birds, themeasures of diversity within that class should include statistical estimates of species richness as well asequitability or dominance. One should examine population responses by the very abundant species, veryrare species, and those in between. Tests of habitat indicators for community diversity should accompanytests of habitat indicators for favored species (e.g., ring-necked pheasants, Phasianus colchicus, or graypartridge, Perdix perdix).

The National Resources InventoryMuch of the data in the succeeding pages is the result of analyses of the 1997 National ResourcesInventory (NRI, Natural Resources Conservation Service 2000). The NRI is an inventory of soil, water,land cover/uses and related resources on the nonfederal lands of the U.S. It is a stratified random sampleof over 800,000 points that has been repeated at 5-year intervals since 1982. The purpose is to obtainstatistically reliable estimates of the conservation treatment of the nonfederal lands for use by the U.S.Department of Agriculture, Congress, policy makers, and others in evaluating national programs andpolicies pertaining to agriculture, land use and natural resources occurring on nonfederal lands. In thefollowing pages NRI data are displayed by Farm Resource Regions and by hydrologic units. The FarmResource Regions are defined primarily for economic analyses by the Economics Research Service, U.S.Department of Agriculture. The hydrologic units represent the boundaries of large watersheds defined bythe U.S. Geological Survey. This scale represents the 4-digit hydrologic unit scale although hydrologicunits can be aggregated or divided from this scale for other purposes. It should be noted that neither ofthese categories are “ecological regions” although as progress is made we will move toward ecologicalregions as an analysis framework.

3

Land UseSeventy-nine percent of the land in the United States is in nonfederal ownership. Land in federalownership consists primarily of forests, rangelands, and parks. Some federally owned rangeland is used byprivate individuals to graze livestock on an annual fee basis. However most agricultural production occurson nonfederal land. Major land uses in the United States for 1997 (Natural Resources ConservationService, 2000) are rangeland (20.9%), forest (20.9%), federal (20.7%), cropland (19.4%), pasture (6.2%),urban, built-up and roads (5.0%), Conservation Reserve Program (CRP, 1.7%), and miscellaneous (waterbodies, mined land, barren, etc., 5.2%). Rangeland is land used by grazing animals where the managementconsists of manipulating the vegetation primarily by adjusting grazing extent, or by prescribed fire, andother methods generally without cultivating the soil. Most rangelands would fall under the OECDdefinition of semi-natural agricultural habitats. Pasture is also land used by grazing animals butmanagement consists of planting desired vegetation and applying soil amendments to increaseproductivity. Pastures with low intensity management would be classed as semi-natural agriculturalhabitats while those under a high level of management would be classed as intensively farmed agriculturalhabitats. The CRP is a cropland diversion program where lands with serious soil erosion or relatedproblems were converted from cropland to permanent vegetative cover under 10-year contracts with theUnited States Department of Agriculture. Cropland as used here includes land used to produce allagricultural commodities including grains, seeds, hay, fruits, vegetables, and orchards. Cropland wouldgenerally fall under the OECD definition of intensively farmed agricultural habitats.

The distribution of land uses is primarily determined by landscape characteristics including climate,geology, soil, topographic characteristics, and water. While cropland occurs nearly throughout the UnitedStates it is most abundant in the Heartland (29.2%), Prairie Gateway (19.3%), Northern Great Plains(15.6%), and Northern Crescent (10.5%, Table 1, Figure 1). Rangeland occurs primarily in the arid west-central regions of the Prairie Gateway (32.2%), Basin and Range (23.7%), Fruitful Rim (22.1%), andNorthern Great Plains (20.5%, Table 1, Figure 1). Cropland, pasture, rangeland and other agriculturalfields occur in mosaic patterns with other land uses across the landscape. Their value as habitat, and thevalue of habitat attributes associated with them, are governed by the particular characteristics of thelandscape mosaic in which they occur. Consequently it is difficult to quantify the habitat value ofagricultural fields without including an analysis of the larger landscape mosaic in which they occur. Theconcept of ecological regions will be useful for large-scale analyses. Even at finer scales the value ofhabitat elements on agricultural fields will be primarily defined by the permanent or residual vegetation(i.e., hedgerows, field margins, borders, or buffer strips, riparian features, etc.) remaining on the landscapeand the surrounding land uses.

Land use changes occur as economic conditions and farm needs dictate. Cropland, pasture, and rangelandall exhibited net declines during the period 1992 – 1997 while forest and developed land both increased(Table 1). These net changes are the result of many local changes as land use shifts to and from each landuse category. For example, although the net change in cropland was -2,141,100 hectares this was the resultof converting 17,180,500 hectares of cropland in 1992 to other land uses in 1997, while converting15,039,400 hectares of other land uses to cropland in 1997. Hence the effect on wildlife habitat issubstantial as 32,219,900 hectares were involved in land use changes centered on cropland. The rate withwhich land use changes occur may

4

Table 1. Extent of important land cover/uses and wetlands in 1997 and net changes during 1992 – 1997 by Farm Resource Region (hectares X 1000,Natural Resources Conservation Service, 2000). Margins of error are in parenthesis. The 95% confidence interval is defined as the estimate +/- marginof error. Farm Resource Region (FRR) names are: H - Heartland, NC - Northern Crescent, NGP - Northern Great Plains, PG - Prairie Gateway, EU -Eastern Uplands, SS - Southern Seaboard, FR - Fruitful Rim, BAR - Basin and Range, and MP - Mississippi Portal.

FRR 1997Cropland

NetChangeCropland

1997Pasture

NetChangePasture

1997Range

NetChangeRange

1997Forest

NetChangeForest

1997Wetland

NetChangeWetland

1997Devel-oped

NetChangeDevel-oped

H 44,510.2(251.4)

322.0(90.4)

7,888.6(167.7)

-609.9(90.1)

745.5(62.9)

-14.2(8.7)

9,496.5(164.4)

198.3(42.7)

2,816.0(96.6)

-2.1(4.3)

5,251.7(104.5)

418.8(26.5)

NC 16,022.4(215.4)

-583.9(74.3)

4,845.3(132.9)

-527.2(79.4)

0 0 43,985.1(348.8)

252.0(73.4)

13,326.6(300.8)

-27.1(6.4)

8,666.9(125.4)

942.9(33.7)

NGP 23,744.0(403.9)

317.1(107.5)

2,235.8(173.7)

-31.3(93.1)

33,659.3(671.5)

-221.2(93.0)

1,762.7(139.4)

4.4(39.1)

3,132.7(165.1)

-0.2(4.0)

1,258.0(64.0)

53.4(10.9)

PG 29,371.9(341.3)

-511.9(93.7)

6,004.3(162.2)

79.3(80.8)

52,667.6(656.7)

17.2(97.7)

3,322.6(271.5)

89.6(44.9)

893.9(65.3)

12.1(3.7)

3,920.3(113.2)

331.8(33.6)

EU 4,763.1(128.7)

-145.3(61.7)

10,844.6(177.2)

-542.0(83.5)

959.2(73.5)

-10.6(13.8)

25,733.1(220.6)

103.3(67.6)

736.8(56.6)

1.3(2.7)

3,869.6(102.3)

585.8(29.0)

SS 7,636.9(154.9)

-528.5(63.7)

7,227.9(167.0)

-440.1(80.4)

568.8(57.8)

5.2(14.6)

37,375.8(251.1)

-12.7(82.0)

9,639.9(187.3)

-25.0(6.9)

6,018.9(123.2)

908.7(39.6)

FR 12,078.6(369.2)

-494.5(101.5)

4,123.4(168.1)

-270.2(93.1)

36,237.4(815.1)

-222.9(155.6)

18,696.4(457.7)

-153.9(113.0)

6,767.5(192.0)

-4.7(21.3)

7,199.4(214.2)

880.7(58.9)

BAR 5,990.7(269.6)

-207.2(65.6)

2,591.1(221.2)

45.3(67.5)

38,880.0(1,204.6)

-122.9(153.4)

13,127.8(543.7)

40.8(137.7)

2,028.6(157.2)

-4.7(6.1)

1,966.5(100.8)

207.8(29.7)

MP 8,202.6(138.8)

-308.7(46.5)

2,605.1(94.9)

-55.4(49.9)

112.2(17.5)

1.1(4.7)

10,271.7(144.1)

132.0(41.5)

5,599.0(129.0)

-16.0(6.7)

1,332.8(49.6)

161.2(14.6)

Total 152,320.2(811.1)

-2,141.1(242.3)

48,366.1(497.8)

-2,351.6(242.3)

163,829.9(1,735.1)

-568.4(257.8)

163,771.7 (937.8)

653.9(235.4)

44,941.1(497.7)

-66.4(26.1)

39,484.1(357.4)

4,491.2(100.3)

5

make it difficult for some species to adapt to those changes. Permanent habitat interspersed throughout thelandscape mosaic can have an important dampening effect on short-term land use changes. Vegetativecomposition of natural plant communities, such as forests and rangelands, are complex, diverseassemblages of plant species. Their conversion to agriculture or other land uses represents a substantialloss of biologically diverse habitat. This complexity is not replaced merely by planting but rather itdevelops over long periods of time. The measurement of land use changes between agriculture and otherland uses, and changes within different agricultural land uses can only be accomplished accurately by aninventory or census where specific fields or sample points are tracked repeatedly over time. Merelyreporting the total area of agricultural lands or specific crop types over time fails to document actualdynamics of land use change.

Wetlands represent a valuable ecological resource with very significant habitat values. Their very highbiological value results from their occurrence at the land-water interface where both terrestrial and aquaticorganisms may interact. Wetlands occur on all land uses (e.g., cropland, forestland, etc.). There werealmost 45 million hectares of wetlands in the U.S. in 1997 (Table 1). There was a net loss of 66,400hectares of wetlands during the interval from 1992 – 1997 (Table 1). This net loss consisted of a gross gainof 138,000 hectares of wetlands and a gross loss of 204,400 hectares. Conversion of wetlands todevelopment accounted for 48.8% of the gross wetland loss while losses due to agriculture accounted for26.5% and losses on forest land accounted for 11.9% (Figure 1). It is only since the early 1980’s thatdevelopment has exceeded agriculture as the principal reason for wetland loss. Wetland losses due toagriculture are declining in the U.S., as farmers are becoming more conscious of environmental values andas federal wetland conservation incentives have been implemented. Many of the recent wetland gains arealso the result of farmers creating or restoring wetlands on land previously used for agriculture, a largeportion of which comes under the Wetland Reserve Program of the U.S. Department of Agriculture. Aswith all natural resources, geographic variation is apparent in wetland abundance, and the reasons forwetland losses. Note from Table 1 that there were “no net wetland changes” in five of the Farm ResourceRegions (i.e., Heartland, Northern Great Plains, Eastern Uplands, Fruitful Rim, and the Basin and Range).The margins of error in those five regions were larger than the estimates; hence the 95% confidenceinterval includes both positive and negative numbers.

The rate with which man induced activities can change the landscape or land uses can exceed the rate thatecological communities can adapt to those changes. Consequently we feel that it is imperative to use theland such that human derived services are within the capability of the land to preserve the integrity,stability, and beauty of the biotic community.

Spatial PatternsSpatial patterns are also important determinants of the ecological value of agricultural habitats. Wildlifeare rarely distributed uniformly in space. Species exhibit a spatial structure in their occurrence andabundance that is caused by habitat heterogeneity, natural disturbance patterns, land use, and resourcemanagement activities. Questions concerning the likely persistence of populations within a network ofhabitat patches, or questions about the most appropriate layout and schedule for land management actionsare now commonplace (Flather et al. in press). In order to account for this spatial complexity, the 1997NRI sampled spatial patterns along X-shaped sets of transects centered over each sample point where eachof the 4 transects was 152.4 meters long. Nine general cover types were interpreted from aerialphotographs of each sample point. From these data several indices of spatial pattern can be developed.The mean length of the first cover type along each of the 4 transects is an indicator of the patch size of thecover type. The size of habitat patches can be of critical importance for some species. As expected thelargest cropland patch sizes generally also occur where cropland is most abundant (Figure 2). Smallercropland patches interspersed among other cover types such as forest or rangeland generally providegreater habitat values for many species, however large cropland fields may attract large flocks of migrating

6

birds for feeding areas during migration. The indicator of rangeland patch size (Figure 3) reveals thatrangelands generally occur in larger patches than cropland.

Figure 1. Important reasons for wetland loss during 1992-1997 by Farm Resource Region (NaturalResources Conservation Service, 2000). Losses displayed represent gross losses while gross gains and netchanges are not displayed in this figure. Farm Resource Region names are: H - Heartland, NC - NorthernCrescent, NGP - Northern Great Plains, PG - Prairie Gateway, EU - Eastern Uplands, SS - SouthernSeaboard, FR - Fruitful Rim, BAR - Basin and Range, and MP - Mississippi Portal.

FR

FR

FRFR

FREU

EU

NGP

PG MP

SSSS

NC

BAR

NC

H

U.S. Totals

35.9%12.2%

9.2%

42.8%

23.0%

19.1%47.7%

10.2%

49.2%

10.7%

2.8% 37.3%

46.5%

1.6%

51.9%

70.9%

14.6%

6.6%

7.9%

14.4%18.1%

1.0%

66.5%

59.6%

17.7%

4.3%

18.4%

39.5%

8.3%

31.7%

20.5%29.6%4.9%

65.5%

11.9%Forest

26.5%Agriculture

12.8%Miscellaneous

48.8%Development

An index of fragmentation was defined as the number of cover type segments per unit of length, orsegments per unit. Cropland cover types are much more highly fragmented in the eastern U.S. thanelsewhere (Figure 4), corresponding also to the areas where the indicator of cropland patch sizes wassmaller. Fragmentation of rangelands was also greatest where rangeland patch size was the smallest(Figure 5).

Many wildlife species will utilize multiple habitats within their daily home range or to complete their lifecycles. For example waterfowl and shorebirds may utilize perennial herbaceous vegetation (i.e., grassesand forbs) as nesting cover, then move to wetlands for brood-rearing purposes. They also may feed on

7

insects in hay meadows or grain from agricultural fields. If any one of these components is missing thehabitat may be useless to them. Consequently ecological or landscape context plays an important role indetermining the value of habitats associated with agricultural lands. Correct interpretation of theses spatialpatterns can only be made via empirical studies correlating measures of biodiversity with these spatiallandscape patterns.

Figure 2. Patch size indicator for nonfederal cropland displayed by hydrologic unit (Natural ResourcesConservation Service 2000). The variable SL1 represents the mean radius in meters for cropland sites.Federal land occurs over about 21% of the U.S., predominantly in the west, but is not identified in thismap.

SL180.8 - 108.5108.5 - 121.0121.0 - 129.8> 129.8

8

Figure 3. Patch size indicator for nonfederal rangeland displayed by hydrologic unit (Natural ResourcesConservation Service 2000). The variable SL1 represents the mean radius in meters for rangeland sites.Federal land occurs over about 21% of the U.S., predominantly in the west, but is not identified in thismap.

SL129.9 - 117.0117.0 - 128.3128.3 - 137.2> 137.2

Figure 4. Indicator of fragmentation for nonfederal croplands displayed by hydrologic units (NaturalResources Conservation Service 2000). The variable segments per unit (SPU) is an indicator of landscapefragmentation with lower values representing more homogeneous cropland patches. Federal land occursover about 21% of the U.S., predominantly in the west, but is not identified in this map.

SPU0.001 - 0.00190.0019 - 0.002310.00231 - 0.00273> 0.00273

9

Figure 5. Indicator of fragmentation for nonfederal rangelands displayed by hydrologic units (NaturalResources Conservation Service 2000). The variable segments per unit (SPU) is an indicator of landscapefragmentation with lower values representing more homogeneous rangeland. Federal land occurs overabout 21% of the U.S., predominantly in the west, but is not identified in this map.

SPU0.00117 - 0.001520.00152 - 0.001890.00189 - 0.00228> 0.00228

U.S. Department of Agriculture ProgramsAs noted previously, agricultural land uses are determined to a large extent by attributes of natural features,including the soil, water, climate, and landscape, then to a lesser extent by policy. Economic conditionsand analysis of the 1982 NRI data prompted federal agricultural policy makers to evaluate cropland use inrelation to soil erosion and land capability factors. Consequently a shift in agricultural policy since 1985has discouraged producing annually cultivated crops on lands where the risks of increased soil erosionwere greater. As a result, soil erosion rates on cropland and CRP land have declined from an average of6.9 metric tons/hectare in 1982 to 4.3 metric tons/hectare in 1997. In similar fashion, 80.8% of cultivatedcropland now has estimated soil erosion rates less than the “tolerable” soil loss level, while in 1982 only73.4% of cultivated cropland had soil erosion rates that were below the “tolerable” soil loss level. Theterm tolerable here refers to the rate of soil erosion that does not jeopardize the long-term sustainability ofthe soil resource and represents a specific value for each soil map unit component. These points illustratethe environmental benefits of those policy shifts - wildlife and aquatic habitats associated with downstreamreceiving waters and wetlands now receive much less sediment and associated pollutants than they did in1982. While these changes may not fall neatly within the realm of the proposed OECD habitat indicatorcriteria, they certainly improve habitat quality. We believe it is important to recognize and receive creditfor national policies resulting in these small, but perhaps extensive, improvements to habitat condition.The approach to habitat indicators should incorporate all agricultural land and not just the most valuedagro-ecosystems.

The Conservation Reserve Program (briefly described on p. 3) has been one of the most beneficialprograms for wildlife habitat in the U.S. The CRP was passed into law in response to massiveoverproduction of agricultural grains accompanied by low prices and excessive soil erosion rates. Erosionreduction from eroding cropland was the primary criteria for participation in the early years of theprogram. Eligibility criteria were broadened in 1990 to include improving water quality and other

10

environmental goals, while in 1996 establishing wildlife habitat was also added. Overall enrollmentpeaked at 14.7 million hectares in 1993. Nearly 9.5% of the nation’s cropland was enrolled in perennialvegetative cover for 10-year contracts. The combined size of the new wildlife habitats created by the CRPwas twice as large as the National Wildlife Refuge System and all state-owned wildlife areas in thecontiguous 48 states combined (USDA Farm Service Agency Online). Recruitment of grassland nestingbirds, many waterfowl species, and other wildlife have shown improvements, some of which areattributable to the CRP. Wildlife habitat was one of many environmental benefits resulting from the CRP.

Other provisions of U.S. agricultural legislation since 1985 also directly or indirectly improved habitat inagro-ecosystems. The Wildlife Habitat Incentives Program provides technical and financial assistance tofarmers who implement specific habitat management practices on their land. The Wetland ReserveProgram provides technical assistance and long-term rental payments to farmers who create or restorewetlands in agricultural landscapes. The Environmental Quality Incentives Program also provides habitatand water quality improvements by applying conservation practices on working lands and reducing soilerosion rates. The impact of these programs on wildlife habitat is described in Heard et al. 2000.

Formal efforts to manage U.S. farmland for wildlife habitat and other natural resource benefits go back tothe 1930s when the Soil Conservation Service (name was changed to the Natural Resources ConservationService in 1994) was established to combat the “Dust Bowl”. Technical assistance and guidance wereprovided to farmers for conservation of all natural resources. Wildlife habitat did not have an economicvalue that could be easily traded in the market place so it was often a secondary or lower managementconcern. Nonetheless many individual farmers have implemented wildlife habitat management practicesas secondary uses of the land behind agricultural production. Often, practices installed to reduce soilerosion can be applied in such manner as to add habitat elements to agricultural lands (Brady 1985). Morecomprehensive habitat management practices can also be installed and integrated into the overallagricultural land management scheme (Warner and Brady 1994). It is important to note that many of thesepractices may not be obvious to the casual observer nor would they necessarily be easy to quantify. Forexample the combination of conservation tillage, grass back-sloped terraces, grass border strips (i.e., fieldmargins), and contour strip cropping would improve habitat values on a farm without reducing economicproduction. Wildlife habitat management at this level is not intensive, but may be extensive. Yet most ofthese practices would go undetected or at least without receiving credit as part of the OECD habitatindicators proposal.

The beneficial effects of habitat elements occurring on agricultural fields (e.g., field margins of beneficialperennial vegetation, hedgerows, etc.) are directly dependent upon the landscape setting, particularecological region, and intensity of land uses. Studies of bird communities indicate that different measuresof bird diversity respond differently to land use and land cover patterns leading us to conclude thatmultiple measures of wildlife community structure should be examined in assessing impacts from landintensification (Brady and Flather 1995). Empirical analyses of wildlife abundance with habitat attributesoccurring on agricultural fields are confounded by other habitat attributes occurring across the landscapematrix. Consequently it is important to conduct empirical studies using measures of biodiversity with thefull suite of landscape attributes. Perhaps from such analyses meaningful indicators of habitat will beidentified.

11

Discussion Items (19 July 2001) “Ecosystem/Habitats Impacted by Agricultural Activities”From the preceding discussion we submit that the approach to habitat indicators should incorporate allagricultural lands rather than the most valued agro-ecosystems. As described above, substantial area ofU.S. cropland has been removed from production and is serving as a “conservation reserve” whileproducing substantial wildlife habitat and other environmental benefits. It is important not to ignore thissignificant contribution. Relatively high value crops occur in nearly every region because of the diversityof specialty crops. Attributes of the landscape matrix within which agro-ecosystems occur should beevaluated jointly with agricultural fields. Therefore all field margins and not just field margins for arablecrops should be included in the analysis. Again suggesting a landscape-level approach to thequantification of habitat indicators. We demonstrated one technique using NRI data for documentingspatial patterns, but other techniques including Geographic Information Systems can be useful. Many ofthese techniques are either labor, capital, or technologically intensive and may not be available in allcountries. However at the very least natural resource and/or land use inventories or censuses should beconducted to monitor changes across the agricultural landscapes. Merely reporting the area in agriculturalproduction yields little information regarding habitat values. Special programs that impact agriculturalhabitats such as the Conservation Reserve Program, Wetland Reserve Program, as well as subtle habitatmanagement efforts occurring secondarily on working agricultural landscapes are also worthy of reporting.

The concept of habitat matrices is good, but as with other indicators of habitat, they may be poor predictorsof species occurrence or biodiversity. Williams et al. (1997) reported that matrices of vegetative covertypes were of marginal value in representing patterns of vertebrate biodiversity, and that ecologicalattributes other than vegetative cover types were more important in recognizing patterns of speciesoccurrence. Their results suggested that the level of detail in a classification of vegetative cover couldinfluence patterns of association exhibited by vertebrate species. Likewise because of the inherentvariability of natural systems one must be careful to separate statistical artifact (i.e., noise) from trueecological patterns.

ConclusionWildlife habitat relationships are very complex and at best we can say that we have developed theoreticalframeworks against which we can test hypotheses about how we think these relationships work. Predictivemodels based upon empirical data about species-habitat relationships have proven useful for landmanagers. However they are more apt to confirm to us what we know about the area we have studied thanto accurately yield a biodiversity profile of similar habitats that as yet have not been inventoried (i.e.,models tend to lack generality). This is not to say that we should abandon the task before us, but ratherthe advancement of the state-of-the-art regarding habitat indicators must be undertaken with scientificrigor. As habitat indicators are selected and evaluated they must be tested against empirically derivedpatterns of biodiversity, including even the basic verification of presence/absence of a species on alandscape. This is an iterative process based upon continually refining the store of our knowledge.

Literature Cited:

Brady, S. J (1985). Important soil conservation techniques that benefit wildlife. Pages 55-62 inTechnologies to Benefit Agriculture and Wildlife – Workshop Proceedings (Washington, DC: USCongress, Office of Technology Assessment, OTA-BP-F-34, May 1985).

Brady, S. J. and C. H. Flather (1995). Using the National Resources Inventory in wildlife assessmentmodels. Pages 92-101 in Lee, L. K. Proceedings of the 1992 National Resources InventoryEnvironmental and Resource Assessment Symposium. U.S. Department of Agriculture, NaturalResources Conservation Service, Washington, DC.

12

Flather, C. H., M. Bevers, J. Hof. In Press. Prescribing habitat layouts: analysis of optimal placement forlandscape planning. Pages 000-000 in K. J. Gutzwiller, ed. Concepts and applications forlandscape ecology in biological conservation. Springer-Verlag, New York, New York.

Heard, L. P., A. W. Allen, L. B. Best, S. J. Brady, W. Burger, A. J. Esser, E. Hackett, D. H. Johnson, R. L.Pederson, R. E. Reynolds, C. Rewa, M. R. Ryan, R. T. Molleur, and P. Buck (2000). A comprehensive

review of farm bill contributions to wildlife conservation, 1985-2000. W. L. Hohman and D. J.Halloum, editors, U.S. Department of Agriculture, Natural Resources Conservation Service,Wildlife Habitat Management Institute, Technical Report, USDA/NRCS/WHMI-2000. 208 pages.

Leopold, Aldo (1933). Game Management. Charles Scribner’s Sons, New York, New York. 481 pages.

Leopold, Aldo (1966). A Sand County almanac with essays on conservation from Round River. OxfordUniversity Press, New York, New York. 295 pages.

Natural Resources Conservation Service (2000). 1997 National Resources Inventory. U.S. Department ofAgriculture, Natural Resources Conservation Service. http://www.nhq.nrcs.usda.gov/NRI.

Schamberger, M. 1988. Monitoring wildlife habitat – a critique of approaches. Statistical Journal of theUnited Nations ECE 5:303-313.

USDA Farm Service Agnecy Online. The Conservation Reserve Program.http://www.fsa.usda.gov/dafp.cepd/12crplogo/tableof.htm. 39 pages.

Van Horne, B. 1983. Density as a misleading indicator of habitat quality. J. Wildlife Manage. 47(4):893-901.

Warner, R. E. and S. J. Brady (1994). Managing farmlands for wildlife. Pages 648-662 in T. A. Bookhout(ed) Research and Management Techniques for Wildlife and Habitats. The Wildlife Society,Bethesda, Maryland.

Williams, B. K., S. Mahabir, J. Schlagel, and D. E. Capen. 1997. Patterns in wildlife-vegetationassociation matrices. Journal of Environmental Management 51:1-13.