conservation implications of climate change: soil …precipitation;two primary conservation...

CONSERVATION IMPLICATIONS OF

CLIMATE CHANGE:

SOIL EROSION AND RUNOFF FROM CROPLAND

A Report from the Soil and Water Conservation Society

®


CLIMATE CHANGE:

SOIL EROSION AND RUNOFF FROM CROPLAND

A Report from the Soil and Water Conservation Society

January 2003

CONSERVATION IMPLICATIONS

OF CLIMATE CHANGE: SOIL EROSION

AND RUNOFF FROM CROPLAND

A report from the Soil and Water Conservation Society, withfinancial support from the U.S. Department of Agriculture’sNatural Resources Conservation Service, Cooperative AgreementNo. 68-3A75-2-98.

Copyright © 2003 by the Soil and Water Conservation SocietyAll rights reservedManufactured in the United States of AmericaPrinted on recycled paper

The Soil and Water Conservation Society (SWCS) is a nonprofitscientific and educational organization that serves as an advocatefor conservation professionals and for science-based conservationpolicy. SWCS seeks to advance the science and art of soil, water,and related natural resource conservation to achieve sustainability.Members practice and promote an ethic that recognizes the inter-dependence of people and their environment.

SWCS has about 10,000 members around the world. Theyinclude researchers, administrators, planners, policymakers, teach-ers, students, farmers, and ranchers. Nearly every academic disci-pline and many different conservation institutions are representedwithin the membership.

Member benefits include the widely respected Journal of Soil andWater Conservation; representation in policy circles; opportunitiesfor leadership and networking; and discounts on books and con-ference registrations.

SWCS chapters throughout the United States, Canada, and theCaribbean Basin conduct a variety of activities at local, state, andprovincial levels and on university campuses. These 75 chaptersrepresent the grassroots element of the organization. Each chapterelects its own officers, organizes conservation forums, and formu-lates local recommendations on land and water conservation issues.

Soil and Water Conservation Society945 Southwest Ankeny RoadAnkeny, Iowa 50021

Telephone: (515) 289-2331Fax: (515) 289-1227www.swcs.org

4 Expert Panel

5 Executive Summary

6 Introduction

7 Soil Erosion, Runoff, and Precipitation

8 Simulated and Observed Changes in Precipitation

Simulated Changes in Precipitation Regime

Observed Changes in Precipitation Regime

Key Findings

14 Estimated Effects on Soil Erosion and Runoff

Estimated Effects on Soil Erosion

Estimated Effects on Surface Runoff

Key Findings

17 Conservation Implications

Conservationists Should Be Concerned

How Should Conservationists Respond?

Update Climatic Parameters

Improve Damage Estimates

Conservation Planning as Risk Management

21 References

23 Appendix

CONTENTS

CONSERVATION IMPLICATIONS OF CLIMATE CHANGE 3

SOIL AND WATER CONSERVATION SOCIETY4

Jeri Berc

Natural Resources Conservation ServiceU.S. Department of Agriculture

Jim Bruce

Soil and Water Conservation Society

Dave Easterling

National Climatic Data Center

P.Y. Groisman

National Climatic Data Center

Jerry Hatfield

National Soil Tilth LaboratoryAgricultural Research ServiceU.S. Department of Agriculture

Bill Hughey


Greg Johnson

Water and Climate CenterNatural Resources Conservation ServiceU.S. Department of Agriculture

Bob Kellogg


Rick Lawford

Office of Global ProgramsNational Oceanic and Atmospheric Administration

Linda Mearns

National Center for Atmospheric Research

Richard Moss

Office of the U.S. Global Change Research Program

Mark Nearing

National Soil Erosion Research LaboratoryAgricultural Research ServiceU.S. Department of Agriculture

Monte O’Neal

National Soil Erosion Research LaboratoryAgricultural Research ServiceU.S. Department of Agriculture

Ken Spaeth

Agricultural Research ServiceU.S. Department of Agriculture

Conservation Implications of Climate Change

EXPERT PANEL


EXECUTIVE SUMMARYA changing climate will affect soil and water resources on agri-cultural land in many ways, but will the effect of climate changeon soil and water resources on agricultural land be large enoughto warrant changes in U.S. conservation policy or practice? Toanswer this question, the Soil and Water Conservation Society(SWCS) reviewed the literature and engaged members of anexpert panel in a discussion of quantitative estimates of theeffects of climate change on soil and water resources in agri-cultural landscapes. We chose to focus on one climatic variable,precipitation; two primary conservation effects, soil erosion andrunoff; and one type of agricultural land, cropland.

In the end, our answer to this question was an emphatic yes.Conservationists should be seriously concerned about theimplications of climate change—as expressed by changes inprecipitation patterns—for the conservation of soil and waterresources in the United States. The magnitude and extent ofincreased rates of soil erosion and runoff that could occur undersimulated future precipitation regimes are large. More impor-tantly, analyses of the climate record in the United States show


CLIMATE CHANGE:SOIL EROSIONAND RUNOFF FROM CROPLAND

A Report from the Soil and Water Conservation SocietyJanuary 2003

that changes in precipitation patterns are occurring now. Infact, the magnitude of observed trends in precipitation and thebias toward more extreme precipitation events are, in somecases, larger than simulated by global climate change models,particularly since 1970. Extrapolating those relationships to thechanges in precipitation observed over the past century suggestsincreases in soil erosion ranging from 4 percent to 95 percentand increases in runoff from 6 percent to 100 percent couldalready be evident on cropland in some locations.

Unless additional protective measures are taken, such increas-es in soil erosion and runoff from cropland—if widespread—could reverse much of the progress that has been made in reduc-ing soil degradation and water pollution from cropland in theUnited States.

The potential for climate change—as expressed in changedprecipitation regimes—to increase the risk of soil erosion, surfacerunoff, and related environmental consequences is clear. Theactual damage that would result from such a change is unclear.Regional, seasonal, and temporal variability in precipitation islarge both in simulated climate regimes and in the existing cli-mate record. Different landscapes vary greatly in their vulnera-bility to soil erosion and runoff. Timing of agricultural produc-tion practices creates even greater vulnerabilities to soil erosionand runoff during certain seasons. The effect of a particularstorm event depends on the moisture content of the soil beforethe storm starts. These interactions between precipitation, land-scape, and management mean the actual outcomes of any partic-ular change in precipitation regime will be complex.

In sum, a change in precipitation regime also produces achange in the level of risk to which agricultural land is exposed.In general, a regime with greater annual precipitation—partic-ularly if increased storm intensity changes more than storm fre-quency—heightens the risk of soil erosion, runoff, and relatedenvironmental and ecological damages. In general, the riskincreases at a greater rate than precipitation amount or intensi-ty increases. Whether that new, more risky baseline conditiontranslates into greater soil degradation, pollution of surfacewater, pollution of groundwater, or a combination of all threeoutcomes is highly dependent on other factors.

We identified three particularly promising approaches tobegin adapting soil and water conservation policies and prac-tices to a changing precipitation regime.


1. Immediately update climatic parameters in critical conservationplanning tools. Two key efforts already are underway. TheNatural Resources Conservation Service (NRCS) is currentlyupdating the climate components of its primary erosion pre-diction and conservation planning tool using precipitation andtemperature data covering the period 1971-1999.The NationalOceanic and Atmospheric Administration’s National WeatherService (NOAA-NWS) Hydrometerological Design StudiesCenter is updating precipitation frequency studies used in engi-neering applications, with support from NRCS, the U.S.ArmyCorps of Engineers, and others. These two efforts should becompleted as soon as possible and then updated more fre-quently in the future.

2. Undertake targeted investigations to firm up estimates of the dam-age that would likely occur under simulated or observed climate regimes.Those investigations should include the following: (1) replicat-ing studies at more locations and more land uses to develop bet-ter estimates of the variability in soil erosion and runoffresponses to an altered precipitation regime; (2) utilizing theNRCS National Resources Inventory framework to updatenational and regional estimates of soil erosion, runoff, and relat-ed environmental outcomes based on updated climate parame-ters in the Universal Soil Loss Equation or the RevisedUniversal Soil Loss Equation; (3) utilizing the NRCS NationalResources Inventory framework to conduct sensitivity analysesand simulate the effect of simulated climate scenarios onnational and regional estimates of soil erosion, runoff, and relat-ed environmental outcomes; and (4) completing targeted eval-uations and studies to better understand the causes of and future

trends in changes in precipitation regime, particularly thoseportions of the regime of most concern for soil erosion, runoff,and related environmental effects.

3. Evaluate the benefits of building the risk of damage from severerainstorms into the conservation planning process through risk-basedassessments targeted to key conservation systems and environmentaloutcomes. In current practice, conservation systems applied toagricultural land are usually designed, planned, and implement-ed to address soil erosion, runoff, and related effects under esti-mates of expected average annual climate regimes.Conservation planning approaches based on annual averageprecipitation should be reevaluated in light of the evidence thatfuture—or current—climate regimes are characterized byincreased frequency of extreme events. Such a change in con-servation planning would require (1) identifying and quantify-ing thresholds of soil erosion, runoff, or transport of pollutantsabove which damage at some specified frequency is unaccept-able; (2) identifying critical times during the year when thenature and timing of agricultural production practices increasethe risk of damage; (3) determining the probability of occur-rence of storms of sufficient intensity and timing to cause dam-age that exceeds threshold levels; and (4) designing conservationsystems to be protective at vulnerable times and during partic-ular events.

Conservationists should take these three foregoing stepsquickly. Policymakers and program managers need answerssoon so they can begin to make the investment decisions andchanges in operational policy needed to ensure U.S. soil andwater resources are protected under a changing climate.


INTRODUCTIONA changing climate will affect soil and waterresources on agricultural land in many ways. Willthe effect of climate change on soil and waterresources on agricultural land be large enough towarrant changes in U.S. conservation policy orpractice? To answer this question, we reviewedthe literature and engaged members of an expertpanel in a discussion of quantitative estimates ofthe effects of climate change on soil and waterresources in agricultural landscapes.

Climate change will affect soil and water con-servation through multiple pathways becausemany climatic variables have important effects onconservation outcomes. Those variables includeprecipitation, temperature, wind, solar radiation,and atmospheric carbon dioxide, among others.Change in any single variable also is complex. Achange in temperature, for example, will affectconservation differently if that change primarilyaffects minimum, maximum, or mean tempera-ture. A change in a climatic variable also maydiffer seasonally or geographically.

The interaction between and among cli-matic variables and conservation outcomes isdynamic and often nonlinear. Climatic vari-ables interact to magnify or dampen conser-vation effects. Likewise, conservation effectsfeed back into the system and modify theinfluence of climatic variables.

Those interactions could have profoundeffects on soil, water, and related naturalresources. Water budgets, streamflow, and fre-quency and severity of floods and droughtsmay be altered. Biotic communities, plantgrowth and development, and land use pat-terns may shift. Those changes, in turn, mayhave important implications for soil,water, andair quality, as well as fish and wildlife habitat.

From the outset of the project, it was clearthat we needed to simplify the problem byfocusing only on a few of the multiple fac-tors, interactions, and potential conservationoutcomes. Ultimately, we chose to focus onone climatic variable, precipitation; two pri-mary conservation effects, soil erosion andrunoff; and one type of agricultural land,cropland. More specifically, we attempted tosummarize quantitative estimates of theeffects that changes in precipitation amounts,frequency, and intensity might be expected tohave on rates of soil erosion and volumes ofrunoff from cropland. Even with this limitedfocus, the interactions among variables andeffects are numerous (Figure 1).

We elected to focus our work on the effects ofchanges in precipitation on soil erosion andrunoff from cropland because (1) quantitativeestimates were available in the published literaturefor these factors, (2) erosion and runoff areamong the most important factors influenc-

Figure 1Effects of climate change on soil and water resources.

Precipitation

CO2

Temperature

Solar Radiation

Soil Moisture

Biomass Management

Water Supply

Stream Flow

Channel Stability

Flooding

WaterPollution

Soil Degradation Water Pollution

EROSION

RUNOFF(Leaching)

ing agriculture’s effects on soil and waterresources, and (3) scientific understanding ofthe processes underlying the conservationimplications of climate change suggest thatprecipitation would be a dominant factoraffecting conservation outcomes and that soilerosion and runoff would be particularly sen-sitive to changes in precipitation.

We approached our work in three steps. First,we sought to determine the range of effectsthat climate change might have on precipita-tion regimes. We did this by looking at thework of various climate change assessmentprojects—primarily those conducted by theIntergovernmental Panel on Climate Change(IPCC) and the U.S. National AssessmentSynthesis Team (U.S. NAST)—for indicationsof the large-scale changes in precipitationregimes that model simulations and relatedanalyses have produced. We also reviewed, insome detail, recent analyses of observedchanges in precipitation regimes in the UnitedStates during the 20th century.

Our second step was to review publishedstudies that would help us assess, quantitative-ly, the changes in rates of soil erosion andrunoff that could occur under such changedprecipitation regimes.

Finally we relied on the professional judgmentof our expert panel and Soil and WaterConservation Society (SWCS) staff to answerour primary question: Are the effects ofchanging precipitation regimes on soil ero-sion and runoff from cropland likely to belarge enough to warrant changes in U.S. con-servation policy or practice?

Indicates the pathways and effectsinvestigated in this report.


SOIL EROSION, RUNOFF,AND PRECIPITATIONSoil erosion by water remains among themost important forces of soil degradation inthe United States and worldwide. Soil ero-sion on cropland was reduced in the UnitedStates by 40 percent between 1985 and 1995,but progress has stalled since. More than 26million hectares (65 million acres) of croplandin the United States continue to erode at ratesgreater than the soil loss tolerance,T—a stan-dard generally interpreted to indicate that soildepth is being reduced at a rate that threatensthe soil’s agricultural productivity (U.S.Department of Agriculture, 1997). Otheressential environmental and ecological func-tions of soil are degraded by soil erosion, butno national estimates exist of the extent towhich those functions are being impaired.Accelerated erosion is the dominant cause ofsoil degradation globally;water erosion affectsan estimated 1,094 million hectares (2,706million acres) (Oldeman, 1994).

Soil erosion and runoff from croplandaffects water resources directly by deliveringsediment,pollutants attached to sediment, andpollutants in solution to surface water; indi-rect effects occur through changes in streamchannel dynamics and watershed functions.In the United States, nonpoint-source pollu-tion is the single largest cause of surface waterpollution. Soil erosion and runoff from agri-cultural land are major causes and, in manywatersheds, the most important causes ofnonpoint-source pollution. In addition, thepartitioning of precipitation between runoffand infiltration has profound effects on waterbudgets in watersheds. Runoff volumes, peakflow rates, and timing can dramatically affectstream channels, water supplies, and down-stream flooding.

Soil erosion and runoff from cropland willrespond to changes in climate for a number ofreasons, including climate-induced changes inplant biomass, plant residue decompositionrates, soil microbial activity, evapotranspira-tion rates, soil surface crusting and sealing,and shifts in land use that occur as adjust-ments to a changed climate (Williams et al.,1996; as cited by Nearing, 2001). The mostimportant effect of climate change on soilerosion and surface runoff, however, willcome from climate-induced changes in thevolume and erosive power of rainfall(Nearing, 2001).

The amount and intensity of precipitation

affect the amount of soil that can be erodedand the volume and flow rate of surfacerunoff. As the accumulated quantity of rain-fall increases during a precipitation event, thecapacity of the soil to absorb the precipitation(infiltration rate) decreases and surface runoffensues. Initially, runoff water flows uniform-ly across the surface; soon, however, runoffforms roughly parallel rills that coalesce intosmall channels of concentrated flow,which, inturn, coalesce into larger channels, eventuallyentering a stream. Runoff volume,depth, andflow velocity all increase with increasing vol-ume and intensity of precipitation.

Erosion is the process of soil particledetachment and movement. Water flowingacross the soil surface—surface runoff—is themost important detachment and transportmechanism. The erosive power of surfacerunoff is determined by depth of flow, flowvelocity, and the number and energy of parti-cles flowing with the water. The capacity forsurface runoff to carry soil particles increasesas flow depth and velocity increases; thiscapacity, therefore, increases as the volumeand intensity of precipitation increase.Indeed, rainfall erosivity is strongly correlatedto the product of total rainstorm energy andthe maximum 30-minute rainfall intensityduring a storm. Wischmeier and Smithderived this relationship in 1978 (Wischmeierand Smith, 1978). The relationship hasproved to be robust and is still used today inthe soil erosion prediction equations that arethe foundation for erosion control planningin the United States (Nearing, 2001).

SIMULATED ANDOBSERVED CHANGES INPRECIPITATIONOur first step toward achieving the objectivesof this project was to determine the range ofeffects that climate change might have onprecipitation regimes. We did this by lookingat the work of two climate change assessmentprojects—those conducted by theIntergovernmental Panel on Climate Change(IPCC) and the U.S. National AssessmentSynthesis Team (U.S.NAST)—for indicationsof the large-scale changes in precipitationregimes that model simulations and relatedanalyses have produced. Then we reviewed insome detail recent analyses of observedchanges in precipitation regimes in theUnited States during the 20th century.

Simulated Changes In PrecipitationRegimeA great deal of scientific effort has beendirected at understanding the effects ofincreased concentrations of greenhouse gaseson a number of climatic variables. The resultsof that effort have been summarized and syn-thesized in two documents of particularinterest to this project: “Climate Change2001: The Scientific Basis,” prepared byWorking Group I of the IntergovernmentalPanel on Climate Change (IPCC WorkingGroup I, 2001a), and “Climate ChangeImpacts on the United States,” prepared bythe U.S. National Assessment Synthesis Team(U.S. NAST, 2001).

The review by the IPCC working groupconcluded that the following effects were“virtually certain or very likely” to occur inresponse to increased concentrations ofgreenhouse gases in the atmosphere:

• Globally averaged mean water vapor,evaporation, and precipitation increase.

• Mean precipitation increases in mosttropical areas; mean precipitationdecreases in most sub-tropical areas; andmean precipitation increases in the highlatitudes.

• Intensity of rainfall events increases.• There is a general drying of the mid-

continental areas during summer(decreases in soil moisture).

In addition, the working group concludedthat new results indicate it is “likely” that precip-itation extremes increase more than precipitationmeans and the return period for extreme pre-cipitation events decreases almost everywhere.

The magnitude of the simulated changes in


Erosion, Runoff, and Water Pollution

There are four major pathways through whichpollutants are delivered to surface water andgroundwater from cropland. Sediment and sedi-ment-adsorbed pollutants, such as phosphorus,ammonium, and strongly adsorbed pesticides, canbe transported to surface water by erosion. Solublepollutants, such as nitrates, soluble phosphorus,and highly soluble pesticides, can be transported tosurface water dissolved in runoff water. Solublepollutants also can be transported to surface waterin subsurface flow through the soil profile. Finally,soluble pollutants can be transported to groundwa-ter through subsurface flow to underlying aquifers.

The pathway precipitation follows through thesoil system will significantly affect the quantitiesand types of pollutants delivered to surface waterand groundwater. Partitioning of precipitationbetween infiltration into the soil and surface runoffis the most important factor in determining which ofthe four pathways dominates during a storm event.

Infiltration is the process of water entering thesoil during a rainstorm event, snowmelt, or irriga-tion. Infiltration rate (quantity/unit time) is influ-enced by many soil physical and chemical proper-ties. Under most conditions, however, infiltrationrate decreases as the soil becomes more saturat-ed. If the soil is dry, water is absorbed inside soilaggregates or adsorbed onto the surface of soil par-ticles. As soil is wetted by rainfall, the cohesiveforces (energy holding soil particles together) of thesoil become less and less. As the soil approachesits saturation point, cohesive forces reach a mini-mum. Under these conditions, raindrop energy isdissipated on the soil surface, causing dislodge-ment and splashing of soil particles. Thesesplashed soil particles return to the soil surface tofill in the low areas of the soil relief and plug soilpores, slowing infiltration. As the soil pores areplugged, forming a soil crust, infiltration ratedecreases and soil cohesive forces again begin toincrease. Soil particle splash decreases becausemore energy is required to dislodge the soil parti-cles. Surface runoff increases proportionally to thedecrease in infiltration rate.

Water is attracted to soil particles because ofthe di-polar nature of the water molecule. Thisattractive force of the soil particle on water mole-cules becomes less as the water film thickness on

soil particles increases. Therefore, water can movedownward by gravity more easily from soil particle tosoil particle as soil water content increases. Inlarge soil pores, the water films become so thickthat water is not influenced by the attractive forcesof the soil, and the water flows through verticalpores by gravity. This preferential flow can lead tolarge volumes of water transmitted through the soilprofile rapidly, providing these pores extend to thesoil surface and are continuous to deeper depths.Water flowing through large pores in this fashionalso can carry dissolved pollutants to surface watervia underground tile, or to groundwater.

In general, soil erosion and surface runoff arethe dominant pathways affected by an increase inprecipitation intensity. Delivery of sediment, sedi-ment-absorbed pollutants, and soluble pollutants inrunoff water will increase. The processes of ero-sion and runoff from cropland are described else-where in this paper.

Sediment transported to surface water by ero-sion tends to carry a higher concentration of pollu-tants than is contained in the topsoil generally.This is because the heavier particles of sand andsilt in runoff fall out of suspension more easily thanclay and organic matter particles. Because clay andorganic matter have high specific surface area anda higher proportion of cation exchange sites, erod-ed sediment contains a higher proportion of nutri-ent ions, like phosphate and ammonium, relative tothe bulk soil from which it eroded—a process calledenrichment. Researchers have studied this processand calculated enrichment ratios for various sedi-ment components and soil management conditions(Sharpley et al., 1995; Young et al., 1986). Younget al. 1986 evaluated nutrient enrichment ratiosfrom soil erosion plots at three locations in theUnited States. He concluded that nutrient enrich-ment ratios varied with crop rotation, current crop,tillage system, soil mineralogy, and latitude.

An increase in the number of wet days with noincrease in precipitation intensity will tend toemphasize subsurface flow to surface water andgroundwater. Hatfield et al 1998, for example,found that delivery of nitrate increased linearly withcumulative weekly precipitation in a tile-drainedwatershed in Iowa.

Rainfall

Surface Runoff WaterPlus Sediment

SurfaceWaterResource

Aquifer

Soil Surface

Soil MoistureStorage

Artificial Subsurface Drainage or Base Flow

Deep

Percolation

precipitation differed, depending on whichsimulation was used.Three ensembles of mul-tiple simulations were used to make projec-tions. An ensemble of simulations based onan idealized 1 percent per year compoundincrease of carbon dioxide resulted in projec-tions of a global mean temperature increase of1.1 degree Celsius to 3.1 degrees Celsius,with an average of 1.8 degrees and a standarddeviation of 0.4 degree at the point whencarbon dioxide concentrations doubled after70 years. Use of the same ensemble resultedin projected changes in global mean precipi-tation ranging from –0.2 percent to 5.6 per-cent,with an average of 2.5 percent and a stan-dard deviation of 1.5 percent. The relativeagreement among models using the ensemblewas much higher for projected changes intemperature than for changes in precipitation.

A second ensemble was constructed fromsimulations that use estimates of observedforcing during the 20th century to constructa baseline; future forcing from greenhousegases and sulfate aerosols is projected fromthat baseline. Use of this ensemble projectedan increase in global mean temperature of 1.3degrees Celsius, with a range of 0.8 degree to1.7 degrees. Global average precipitation wasprojected to increase by 1.5 percent, with arange from 0.5 percent to 3.3 percent. Thepoints of comparison were 30-year aver-ages—2021 to 2050 versus 1961 to 1990.

The third ensemble of simulations used asan initial state the end of 20th century inte-grations; new forcing scenarios to the year2100 were then used to project climateeffects. The 30-year average temperatureresponse (2071 to 2100 compared to 1961 to1990) was an increase of 3.0 degrees Celsiusor 2.2 degrees Celsius, depending on whichof two forcing scenarios was used. Averageglobal precipitation was projected to increaseby 3.9 percent, with a range of 1.3 percent to6.9 percent, or 3.3 percent, with a range of1.2 percent to 6.1 percent, again dependingon which forcing scenario was used.

The working group also reported that newresults confirmed projections of precipitationevents with increased intensities in a futureclimate with increased greenhouse gases—afinding that was among the earliest modelresults and remains a consistent result in anumber of regions with improved, moredetailed models. The percentage increase inextreme rainfall is greater than the percentageincrease in mean rainfall, and the return peri-od of extreme precipitation events is short-


Figure 2Projected changes in intensity of U.S. precipitation for the 21st century.

The projected changes in precipitation over the United States as calculated by two models indicate that most of the increase is likely to occur in the

locally heaviest categories of precipitation. Each bar represents the percentage change of precipitation in a different category of storm intensity. For

example, the two bars on the far right indicate that the Canadian Centre model projects an increase of over 20 percent in the 5 percent most intense

rainfall events in each region, whereas the Hadley Centre model projects an increase of more than 55 percent in such events. Because both historic

trends and future projections from many global climate models indicate an increase in the fraction of precipitation occurring during the heaviest

categories of precipitation events in each region, a continuation of this trend is considered likely. Although this does not necessarily translate into an

increase in flooding, higher river flows are likely to be a consequence.

Source: U.S. Department of State, 2002.

60

50

40

30

20

10

0

-10

Tren

ds (

% c

hang

e in

pre

cipi

tati

on p

er c

enut

ry)

0%

Lightest

5%

Moderate

Heaviest

5%

20% 40% 60% 80%100%

Hadley ModelCanadian Model

2.60.1

2.5 4.0 3.1 2.9

-6.0 -6.2 -4.7-6.5

-4.3

-9.0

1.54.7 5.1 4.5

6.5

-10.0

-5.4-6.9

-9.0-5.7

7.9 8.6

-7.7

-3.8

6.58.8

1.33.7

11.6

20.3

57.2

7.96.4

8.310.6

22.8

-3.00.3

itation regimes for the United States that aresimilar to those the IPCC concluded werelikely or very likely to occur at global scales.Both scenarios show increases in temperatureand precipitation over the United States, andboth show an increase in heavy precipitationevents in the United States as the climatewarms (Figure 2).

Observed Changes in PrecipitationRegimeIndications of trends in precipitation patternsin the United States found in the observedclimate record are particularly interestingbecause of their immediate implications forconservation. Several published analyses indi-cate that climate has changed during the pastcentury (Groisman and Easterling, 1994; Karland Knight, 1998; Kunkel et al., 1999a,1999b; Peterson et al., 1995; Brown andBraaten, 1998; Changnon, 1998; Frei et al.,1999; Easterling et al., 2000c; Groisman et al.,

1994, 1999a and 1999b; and Cayan et al.,2001 as reviewed by Groisman et al., 2001).In the contiguous United States, those gener-al trends include:

• Increase in minimum temperature.• Decrease in extent of spring snow cover

in the West.• Increase in near-surface humidity and

cloudiness.• Increase in mean precipitation.• Increase in heavy and very heavy rains in

the East.• Increase in high streamflow events in the

East.The observed changes are consistent with

changes projected by the modeling studiessummarized above. Observed increases inmean precipitation and rainfall intensity inthe United States—climatic factors of partic-ular importance to this project—generallyexceed simulated increases.

Annual precipitation, averaged over the con-

ened almost everywhere. Over NorthAmerica, for example, the 20-year returnperiods are reduced by a factor of two, mean-ing the probability of extreme events doubles(Zwiers and Kharin, 1998; as cited by theIPCC Working Group I, 2001a).

The U.S. NAST used two general circula-tion models (GCMs)—the global climatemodel developed by the Canadian Centre forClimate Modeling and Analysis (CGCM1)and the model developed by the HadleyCentre for Climate Prediction and Researchof the Meteorological Office of the UnitedKingdom (HadCM2)—to create climatechange scenarios that could be used toexplore potential effects of climate change inthe United States. The U.S. assessment wasdesigned to identify vulnerabilities to climatechange rather than predict how climate islikely to change in the United States. The cli-mate change scenarios analyzed by the U.S.NAST, however, produced changes in precip-


Figure 3Linear trends in annual precipitation over the contiguous United States, 1900 to 1998.

Trends%/100 yrs

10

20

40

Source: Groisman et al. 2001. Linear trends (percent/100 years) of annual precipitation (1900-

98) over the contiguous United States (updated from Karl and Knight, 1998). Individual trends

from 1221 U.S. Historical Climatology Network stations (Easterling et al. 1996) have been area-

averaged inside the U.S. climatic divisions (Guttman and Quayle, 1996). Gray dots indicate

increasing trends; black dots indicate decreasing trends.

Table 1. Trend characteristics in share of total precipitation occurring in heavy, very heavy,and extreme daily precipitation events over the contiguous United States, 1910-1999.

Annual Precipitation Contribution to Annual Totals

Precipitation Mean Linear Trend Fraction Relative ChangeValue(mm) Estimate Variance Estimate Variance

(%/10 yr) (%) (%/10 yr) (%)

Annual Total 750 0.6 5* 1.00 — —

Heavy 195 1.7 12* 0.26 1.0 20*

(> 95th percentile)

Very heavy 62 2.5 15* 0.08 1.9 17*

(>99th percentile)

Extreme 12 3.3 11* 0.016 2.7 9*

(> 99.9th percentile)* Statistically significant at the 0.05 or higher level.

Source: P. Y. Groisman, U.S. National Climatic Data Center.

tiguous United States, increased about 6 percentover the period 1910 to 1999 (Table 1). Kunkelet al. 1999a estimated a similar upward lineartrend—about 1.3 percent per decade—in annu-al precipitation over the period 1931 to 1996.Both Groisman et al. 2001 and Kunkel et al.1999a found large regional variation in lineartrends in annual precipitation over the periodsthey studied. Over the period 1900 to 1998,most areas showed increasing trends between 10percent and 40 percent (Figure 3). Some areas,however, showed decreasing trends in annualprecipitation. After about 1970, precipitationtended to remain above the 20th century meanand averaged about 5 percent more over thecontiguous United States than in the previous 70years (Karl et al., 1996).

The general trend toward increased precip-itation varied seasonally. Groisman et al. 2001found that linear trends in mean seasonal pre-cipitation, area weighted over the contiguousUnited States, were 0 in winter, 10 percent inspring, 7 percent in summer, and 15 percentin autumn for the period 1910 to 1996.

Linear trends in heavy or very heavy pre-cipitation were more pronounced than thetrend in total precipitation over the contigu-ous United States. The proportion ofincreased annual precipitation occurring inextreme events grew. Precipitation intensity,in other words, increased more than annualprecipitation over the period. Karl andKnight 1998, for example, reported thatheavy and extreme daily precipitation eventsaccounted for most of the increase in annualprecipitation across the contiguous UnitedStates. This bias toward increased precipita-tion intensity has been reported in manylocations worldwide (Easterling et al., 2000a,2000b, 2000c).

Table 1 shows the linear trend in the shareof total precipitation accounted for by heavy,very heavy, and extreme daily precipitationevents in the contiguous United States.The lin-ear trend in total precipitation over the period1910 to 1999 was 0.6 percent per decade. Thelinear trend in the amount of precipitationoccurring during heavy precipitation events(precipitation greater than the 95th percentile forthe location) was 1.7 percent per decade; forvery heavy events (precipitation greater than the99th percentile for the location), the linear trendwas 2.5 percent per decade; and for extremeevents (precipitation greater than the 99.9th per-centile for the location), the linear trend was 3.3percent per decade.

Data regarding the number of days on

which heavy and very heavy precipitationevents occurred showed a similar trend (Table2a). The linear trend in total number of dayson which precipitation greater than 1 mm(0.04 inch) occurred over the period 1910 to1999 was 0.5 percent per decade; the lineartrend in the number of days on which heavyprecipitation events occurred was 1.5 percentper decade; the linear trend in the number ofdays on which very heavy events occurredwas 2.2 percent per decade. The same pattern

was apparent if heavy and very heavy precip-itation events were defined on the basis of theabsolute value of precipitation occurring onthat day (Table 2b).

Kunkel et al. 1999a found the same upwardtrend in heavier precipitation events definedas 7-day duration events with a 1-year returninterval or a 5-year return interval (Figure 4).The overall trend in 7-day, 1-year occurrenceevents was upward at a rate of about 3 percentper decade over the period 1931 to 1996. The


Table 2b. Trends characteristics in number of days with heavy and very heavy precipitationevents over the contiguous United States, 1910-1999 (absolute value definition).

Days with Precipitation Contribution to total dayswith precipitation above 1mm

Events Mean Linear Trend Fraction Relative Change(days/ yr) Estimate Variance Estimate Variance

(%/10 yr) (%) (%/10 yr) (%)

Total days with 75 0.5 6* 1 — —precipitation above1mm

Heavy 1.4 3.3 30* 0.02 2.8 33*

(above 50.8 mm)

Very Heavy 0.13 4.9 22* 0.002 4.4 21*

(above 101.6 mm)* Statistically significant at the 0.05 or higher level.


Table 2a. Trend characteristics in number of days with heavy and very heavy precipitationevents over the contiguous United States, 1910-1999 (percentile definition).

Days with Precipitation Contribution to total dayswith precipitation above 1mm

Events Mean Linear Trend Fraction Relative Change(days/ yr) Estimate Variance Estimate Variance

(%/10 yr) (%) (%/10 yr) (%)

Total days with 75 0.5 6* 1 — —precipitationabove 1mm

Heavy 4.4 1.5 12* 0.06 1.0 11*

(above 95th

percentile)

Very Heavy 0.88 2.2 14* 0.012 1.7 13*

(above 99th

percentile)

Table 3. Trends in share of total annual precipitation occuring in heavy, very heavy, andextreme daily precipitation events in the contiguous U.S., 1910-1970 versus 1970-1999.

1910 - 1970 1970 - 1999

Precipitation Mean Linear Trend Mean Linear Trend(mm) (mm)

Estimate Variance Estimate Variance(%/10 yr) (%) (%/10 yr) (%)

Total Annual 737 -0.4 1 772 1.2 2precipitation

Heavy 188 -0.1 0 208 4.6 12*

(> 95th percentile)

Very Heavy 59 0.9 1 67 7.2 15*

(> 99th percentile)

Extreme 12 1.5 1 14 14.1 22*

(> 99.9th percentile)* Statistically significant at the 0.05 or higher level.


trend in 7-day, 5-year occurrence events wasgreater—about 4 percent per decade over thesame period. Both trends were statistically sig-nificant at the 0.5 percent level. The precipita-tion contributed in 7-day, 1-year recurrenceevents contributed about one-third of theupward trend in annual precipitation.

Table 3 presents data indicating that thepositive trend in heavy and very heavy pre-cipitation events was much more pro-nounced between 1970 and 1999 than from1910 to 1970. The linear trend in total annu-al precipitation, for example,was –0.4 percentper decade over the period 1910 to 1970, but1.2 percent per decade between 1970 and1999. The linear trend in the proportion ofprecipitation occurring in heavy precipitationevents was –0.1 percent per decade between1910 and 1970, but 4.6 percent per decadebetween 1970 and 1999. The differencesbetween the two periods were more pro-nounced for very heavy and extreme events.The same pattern was apparent if the numberof days with heavy and very heavy precipita-tion events were analyzed, but only trends inheavy and very heavy precipitation eventswere significant at the 0.05 level or higher.

Regional variation within the UnitedStates in the trend toward more intense pre-cipitation also is apparent in the climaterecord. Kunkel et al. 1999a found largeregional variation among trends in 7-day, 1-year recurrence precipitation events (Figure5). The largest upward trends—25 percent tomore than 100 percent relative to the 1931-1996 mean—occurred over the southwesternUnited States and over a broad region fromthe central Great Plains, across the MississippiRiver, and into the southern Great LakesBasin. Only a few climate divisions showeddownward trends. Those divisions were inthe northwestern United States and Florida.Climate division trends for 1-day and 3-dayevents were very similar to those for 7-dayevents with a 1-year recurrence. Regionalvariation also was apparent if the trend in rel-ative contribution to total annual precipita-tion made by differing rainfall intensity class-es was analyzed.

Trends in frequency of heavier precipita-tion events varied seasonally as well as geo-graphically. Kunkel et al. 1999a, for example,reported that 7-day, 1-year recurrence eventsoccurred most frequently in the summer inthe Midwest and Great Lakes Basin, but alsooccurred with some frequency in the springand fall. In the Great Plains, most events


Figure 4Trend in heavy precipitation events in the United States, 1931-1996.

Nationally averaged annual U.S. time series of the number of precipitation events of 7-day

duration exceeding 1-year (dots) and 5-year (diamonds) recurrence intervals.

Source: Kunkel et al. 1999a. Copyright held by American Meteorological Society.

0.0

0.2

0.4

0.6

0.8

1.0

1.2

1.4

1.6

Num

ber

of e

vent

s pe

r st

atio

n

1930 1940 1950 1960 1970 1980 1990 2000

Year

Annual

occurred in the summer, while eventsoccurred more evenly across the seasons inthe Northeast and the South. No climatedivisions showed significant trends duringthe winter. On a national average, 37 per-cent of the 7-day, 1-year recurrence eventsoccurred in the summer, 25 percent in thefall, 21 percent in the spring, and 16 percentin winter.

Global trends in 1-day and multiday heavyprecipitation events showed a similar tenden-cy toward more days with heavy precipitationtotals over the 20th century (Easterling et al.,2000c). Nine regions showed statistically sig-nificant positive trends in the average numberof days with heavy rain—(1) eastern UnitedStates, (2) European part of the formerUSSR, (3) Asian part of Russia, (4) southernCanada, (5) coastal regions of New SouthWales and Victoria, Australia, (6) northernJapan, (7) southwestern South Africa, (8)Natal, South Africa, and (9) and Nord-Este,Brazil. Four of those regions showed statisti-cally significant increases in heavy precipita-tion even when total precipitation stayed thesame or decreased. Three regions showed sta-tistically significant negative trends—(1)Ethiopia and Eritrea, (2) equatorial eastAfrica, and (3) Thailand.

Key FindingsUpward trends in total precipitation, coupled with abias toward more extreme precipitation events, areindicated in both simulated and observed climateregimes. The general circulation models usedby the IPCC and U.S. NAST to simulate cli-mate responses to greenhouse gas forcinggenerally agree that an increase in annual pre-cipitation and in the frequency of moreextreme precipitation events are the mostlikely outcomes of climate change resultingfrom increasing greenhouse gas concentra-tions. Analyses of the historic climate recordshow a similar general upward trend in annu-al precipitation and precipitation intensity inthe United States, particularly since 1970.

The difficulties and uncertainties in mod-eling climate response to greenhouse gasforcing are well known and described in thereports of the IPCC 2001a and the U.S.NAST 2001. At the global level, uncertain-ties arise from uncertainties in sensitivities togreenhouse gas emissions and uncertaintiesabout the emissions themselves.

It is possible that the upward trends in pre-cipitation amounts and intensities observed inthe historic climate record will not continue

into the future. It is also possible that thosetrends could accelerate. We did not try toconclude in this project whether changes inthe observed climate record are a result ofanthropogenic forcing; neither did we try toevaluate the likelihood that simulated climatechange in response to greenhouse gas forcingwill actually occur. The concurrence of sim-ulated and observed changes is striking, how-ever, and statistical analyses of the observedclimate record indicate that there is less thana 1 in 1,000 chance that the observed changesin precipitation regime could have occurredunder a stable climate (Karl et al., 1996). TheIPCC Working Group 1 conclusion that“there is new and stronger evidence that mostof the warming observed over the last 50years is attributable to human activities” sug-gests the observed trends could signal thebeginning of a long-term trend (IPCCWorking Group 1, 2001b. p. 10).

Conservationists should be concerned aboutthe potential effects of changes in precipita-tion patterns on soil erosion and runoffregardless of whether those changes areattributed to specific forcings—anthro-pogenic or natural.

Regional, seasonal, and temporal variability inprecipitation is large in both simulated andobserved climate regimes. Regional deviationsfrom simulated global responses are likely tobe quite large, even at fairly large scales.Analyses of the historic climate recordclearly show that variability in change inprecipitation regimes is large. That variabil-ity includes:

• Geographic variability in trends of bothannual precipitation and precipitationintensity.

• Seasonal variability in trends of bothannual precipitation and the frequencyof more extreme precipitation events.


Figure 5Linear trends in frequency of heavy precipitation events over the contiguous United States, 1931

to 1996.

Trends100%50%0%

United States climate division trends in frequency of precipitation events of 7-day duration

exceeding a 1-year recurrence interval. Shaded circles indicate upward trends while open circles

indicate downward trends. The magnitude of the trend is given in terms of the percent increase

or decrease over the period 1931-1996 relative to the 1931-1996 mean. As indicated in the key,

the magnitude of the trend is linearly proportional to the radius of the circle. A tail attached to

the upper right indicates positive trends with local significance at the 5 percent level. A tail

attached to the upper left indicates locally significant negative trends. An “x” indicates a climate

division with no stations with complete records for the period 1931-1996.

Source: Kunkel et al. 1999a. Copyright held by the American Meterological Society.

• Decadal variability in trends of bothannual precipitation and precipitationintensity.

• Geographic variability in the absoluteamounts of rainfall occurring in eventsdefined as heavy, very heavy, or extreme.

Such variability means that general upwardtrends in annual precipitation and precipita-tion intensity will be expressed very differ-ently in different locations, years, and seasons.The risk posed to soil and water resourceswill be just as variable.

Our understanding of the role of climatevariability and forcing factors on precipita-tion is incomplete, but there is clear evidencethat known oscillations have importantimpacts on precipitation regimes. Climatesimulations and analyses of the historic climaterecord suggest a change in baseline conditionscharacterized by a general trend towardincreased total annual precipitation and pre-cipitation intensity. Local and regional varia-tions from that new baseline will be large.

ESTIMATED EFFECTS ONSOIL EROSION ANDRUNOFFOur second step toward achieving the objec-tives of this project was to review publishedstudies that would help us assess, quantitative-ly, the changes in rates of soil erosion andrunoff that could occur under changed pre-cipitation regimes. We focused on studies thatattempted quantitative estimates of change insoil erosion and runoff caused by changes inprecipitation patterns. (Please see the appen-dix for more information about the nature ofthe models used by the authors cited belowto estimate soil erosion and runoff).

Estimated Effects on Soil ErosionSeveral studies have attempted to quantify theeffects of change in precipitation amounts,frequency, and intensity on soil erosion fromcropland.Two studies used output from gen-eral circulation models to estimate the effect

of simulated changes in climate on soil ero-sion rates at the national scale in the UnitedStates. These studies estimated the effect ofclimate change on erosivity by estimating theeffect of a new, simulated precipitation regimeon the R-factor (rainfall-runoff erosivity fac-tor) in the Universal Soil Loss Equation(USLE) or Revised Universal Soil LossEquation (RUSLE).

Nearing 2001 used monthly and annualprecipitation output from the UK HadleyCentre model (HadCM3) and the CanadianCentre for Climate Modeling and Analysismodel (CGCM1) to modify the R factor inRUSLE for the contiguous United States.The period 2080-2099 was compared to thebaseline period 2000-2019. Erosivity was cal-culated from statistically derived relationshipsbetween the R factor and the modifiedFournier coefficient (Renard and Freidmund,1994) and between the R factor and totalaverage annual precipitation. Effects on ero-sivity varied geographically among grid cellsfor which model output was generated.Thecalculated changes in erosivity using theHadCM3 model for individual grid cellsranged from –59.5 percent to +118.3 per-cent. Calculated changes in erosivity amonggrid cells using the CGCM1 model rangedfrom –65.1 percent to +802.8 percent. Theabsolute changes in erosivity averaged over allgrid cells ranged from 15.9 percent to 20.9percent for the HadCM3 model and 53.4percent to 58.3 percent for the CGCM1model, depending on which method wasused to estimate the new R-factors.

Phillips et al. 1993 used four global climatemodels—the National Oceanic andAtmospheric Administration (NOAA)Geophysical Fluid Dynamics Laboratory model(Manabe and Wetherald, 1987), the NASAGoddard Institute for Space Studies model(Hansen et al., 1988), the Oregon StateUniversity model (Schlesinger and Zhao, 1989),and the United Kingdom Meteorological Officemodel (Mitchell et al., 1987)—to generate cli-matic scenarios, assuming a doubling in atmos-pheric carbon dioxide concentrations fromsimulated baseline conditions. Estimates ofannual precipitation from each model wereused to estimate changes in county-averagedUSLE R-factors for each U.S. NationalResource Inventory (NRI) sample point.Model simulations of annual precipitationwere assumed to occur either (1) through achange in storm frequency only or (2)through a change in storm intensity only. In


the first case, a new R-factor was calculatedby multiplying the baseline,county-averaged R-factor by the ratio of simulated-to-baselineannual precipitation. In the second case, a newR-factor was calculated by multiplying the base-line, county-averaged R-factor by the square ofthe ratio of simulated-to-baseline annual precip-itation, adapted from Foster et al. 1985.

Estimated changes in annual soil erosionon cropland, averaged over all U.S. NRI sam-ple points, ranged from 2 percent to 8 per-cent, depending on which model was used ifonly storm frequency changed, and from 5percent to 16 percent, depending on whichmodel was used if only storm intensitychanged. Estimated changes in annual aver-age soil erosion ranged from –2 percent to 10percent on pasture and from –5 percent to 22percent on rangeland, again depending onwhich model was used and whether stormfrequency or storm intensity was adjusted.For all land uses, estimated changes in soilerosion varied greatly from the national aver-ages reported above, depending on the geo-graphic location of the sample points.

Three studies looked at the effects ofchanged precipitation regimes on soil erosionat smaller geographic scales in the UnitedStates. Those studies shed light on the relativeeffects of changes in precipitation frequencyand/or intensity, in addition to changes intotal precipitation.

Savabi et al. 1993 used the Water ErosionPrediction Project (WEPP) model to esti-mate the effect of increased precipitationintensity on soil erosion simulated for crop-land growing corn in Minnesota, Illinois, andTexas. They found that linearly increasing theamount of daily precipitation by 5 percent or10 percent over 50 years increased soil ero-sion by 10.7 percent and 35.6 percent respec-tively. In their study, the amount of precipi-tation received per day (precipitation intensi-ty) was increased while the number of daysreceiving precipitation remained constant.

Pruski and Nearing 2002 compared direct-ly the effects of changes in frequency and/orintensity by allocating a –20, -10,0,10, and 20percent change in annual precipitation tochanges in intensity alone, changes in fre-quency alone, or a combination of changes inintensity and frequency. They simulated theresults of these precipitation regimes usingthe WEPP model for three soil types, threeslopes, and four crops at three locations in theUnited States—West Lafayette, Indiana;Temple, Texas; and Corvallis, Oregon. They

found that a change in intensity had signifi-cantly greater effect on soil erosion than achange in frequency of precipitation.Specifically, a 1 percent change in precipita-tion resulted in an average 2.38 percentchange in soil erosion if precipitation intensi-ty alone was changed. An average 0.85 per-cent change in soil erosion resulted if precip-itation frequency alone was changed.Simulating a combination of change in fre-quency and intensity resulted in an average1.66 percent change in soil erosion for each 1percent change in precipitation.

Lee et al. 1996 used the ErosionProductivity Impact Calculator (EPIC)(Sharpley and Williams, 1990) model to sim-ulate the effect of a change in mean monthlyprecipitation if the frequency of events (num-ber of wet days) or intensity of events(amount of precipitation per day) changed.In contrast to the Pruski and Nearing 2002and Phillips et al. 1993 results, they found lit-tle difference (generally less than 1 percent) inthe mean soil erosion rate resulting fromincreases in frequency compared to increasedintensity of precipitation. They also foundthe mean water erosion rate was linear to pre-cipitation volume for the range of precipita-tion change in their study—a –20 percent to20 percent change in precipitation volumeled to a –37 percent to 37 percent change insoil erosion.They did find, however, that thevariability among the estimates of soil erosionrates resulting from a change in precipitationwas less if the frequency of days receiving pre-cipitation was increased than if storm intensi-ty was increased. This reduced variability wascaused by the effect of the antecedent mois-ture content of the soil. Soils under a precip-itation regime of increased number of wetdays will more likely be wetter and, therefore,more apt to produce soil erosion and runoff.Soils under a precipitation regime of fewerwet days but more intense precipitationevents may or may not produce soil erosionand runoff, depending on the antecedentmoisture status of the soil.

Studies at locations outside the UnitedStates show similar results. Two studies simu-lated the effects of a change in mean month-ly precipitation on erosion rates at the nation-al scale in Great Britain. Boardman andFavis-Mortlock 1993 used the EPIC modelto estimate that increases in mean monthlyprecipitation of 5, 10, and 15 percent fromcurrent conditions in Great Britain wouldresult in 10, 26, and 33 percent increases in

soil erosion respectively. In a similar study,Favis-Mortlock et al. 1991 simulated theeffect of a –10, -5, 5, 10, and 15 percentchange in mean monthly precipitation fromcurrent conditions in Great Britain. Theybased their simulated changes in meanmonthly precipitation on a literature reviewof anticipated climate change by 2050. Theirstudy used the EPIC model to estimatechanges in soil erosion rates ranging from –7percent to 64 percent.

Favis-Mortlock and Guerra 1999 estimat-ed soil erosion changes in Mato Grosso State,Brazil, using precipitation regimes generatedby three different GCMs and a modified ver-sion of the WEPP model (WEPP-CO2 asdescribed by Favis-Mortlock and Savabi,1996). They reported increases in soil erosionranging from 27 percent to 55 percent, basedon model output from two of the GCMs.Output from the third GCM produced a 9percent reduction in soil erosion.

Estimated Effects on Surface RunoffThree of the previously cited studies also esti-mated the effects of precipitation change onsurface runoff. Savabi et al. 1993 reportedthat a 5 percent to 10 percent increase in pre-cipitation intensity (average daily precipita-tion) resulted in a 3.7 percent to 15.9 percentchange in runoff at their three study loca-tions. Lee et al. 1996 reported the same lin-ear relationship between precipitation vol-ume and runoff that was reported betweenprecipitation and soil erosion. A –20 percentto 20 percent increase in precipitation result-ed in an estimated –40 percent to 40 percentchange in runoff.

Pruski and Nearing 2002 found the samerelationship between runoff and precipitationintensity and frequency as they did betweensoil erosion and precipitation intensity andfrequency. They also found that rainfallintensity had a greater effect than rainfall fre-quency on runoff. Each 1 percent change inprecipitation amount resulted in an average2.5 percent change in runoff if a change inintensity accounted for all of the change inamount; an average 1.28 percent change inrunoff occurred if a change in frequencyaccounted for all of the change in precipitationamount;and an average 1.97 percent change inrunoff occurred if a combination of change inintensity and frequency accounted for thechange in precipitation volume.

In addition to the foregoing studies, Nicks1993 used the Simulator for Water Resources


in Rural Basins (SWRRB) model (Arnold etal., 1990) to estimate the effect of changes inprecipitation patterns simulated using the cli-mate generator model (CLIGEN) (Nicks andGander, 1994) on runoff. He reported that a16 percent increase in annual precipitationproduced a 70 percent to 73 percent increasein water yield in three study watersheds incentral Oklahoma under predominatelyrangeland use. A 12 percent increase in pre-cipitation frequency (number of wet days)produced a 47 percent increase in runoff inthe same study watersheds.

Pruski and Nearing (in press) used meanmonthly precipitation outputs from theHadCM3 model to estimate the effect ofchanges in precipitation regime on runoff ineight locations in the United States—Atlanta,Georgia; Cookeville, Tennessee; Corvallis,Oregon; Pierre, South Dakota; Syracuse,Nebraska; Temple, Texas; West Lafayette,Indiana; and Wichita, Kansas. Half of thechange in mean monthly precipitation wassimulated by increasing precipitation frequen-cy (number of wet days) and half by increas-ing precipitation intensity (precipitation perday). They reported that changes in meanannual precipitation ranging from –9.6 per-cent to 10.6 percent produced changes inrunoff from –24.5 percent to 41.0 percent.

Key FindingsChanges in precipitation regimes clearly have thepotential to produce profound effects on soil erosionand runoff from cropland. Simulated changes intotal annual precipitation or mean monthlyprecipitation expressed as either changes instorm frequency, intensity, or some combina-tion of both all produce large changes in esti-mated rates of soil erosion and surface runoff.In all cases, the percentage change in soil ero-sion or surface runoff was larger than the sim-ulated percentage change in precipitation.Soil erosion and surface runoff are highly sen-sitive to changes in precipitation regimes.

An increase in storm intensity—all other factorsbeing equal—will increase soil erosion, runoff, andrelated environmental damage more than anincrease in the number of wet days in a year.Changes in storm intensity have greatereffects on soil erosion and runoff thanchanges in storm frequency. All of thereviewed studies that directly compared theresults of a change in total precipitation allo-cated to either storm frequency or stormintensity—with one exception—reportedlarger effects from changes in storm intensity.

As indicated above, Lee et al. 1996 report-ed a linear relationship between precipitationvolume and soil erosion or runoff with littledifference between a change in storm fre-quency or intensity. The methods used byLee et al. 1996, however, would not beexpected to distinguish effectively betweenthe effects of frequency and intensity. Themodel used—the Erosion ProductivityImpact Calculator (EPIC)—estimates soilerosion using a version of the ModifiedUniversal Soil Loss Equation (MUSLE).Because the model does not contain theGreen-Ampt infiltration model, which esti-mates water infiltration into the soil basedupon the physical condition of the soil andthe rainfall intensity, it is less capable of esti-mating the effects of precipitation intensitythan the physically based models used in theother studies, particularly Pruski and Nearing2002. In addition, runoff in EPIC is calculat-ed using the NRCS runoff curve numbermethod. The curve number method is anempirical formula and cannot account forchanges in rainfall intensity.

Pruski and Nearing used the more physi-cally based Water Erosion Prediction Project(WEPP) model to simulate soil erosion andrunoff. WEPP also contains the Green-Amptinfiltration model. Although both studies con-sider the effects of individual rainfall events onrunoff and erosion, WEPP accounts for thephysical processes of infiltration, soil roughness,and antecedent moisture. Therefore, the Pruskiand Nearing study should give a better indica-tion of the effects of changing rainfall intensityversus rainfall frequency.

In general, increased precipitation intensitymeans a greater proportion of annual precip-itation is apportioned to runoff rather than toinfiltration. Surface transport of pollutants viasoil erosion and runoff will increase marked-ly—all other factors being equal—and surfacetransport of pollutants generally will threatenwater resources more than subsurface trans-port. In addition to the direct threats to waterquality from increased delivery of sedimentand other pollutants (both sediment attachedand in solution) from cropland, increasedstorm intensity will magnify the effects ofrunoff on channel stability and flooding.Increased storm intensity likely would accel-erate both ephemeral and incised gully ero-sion—erosion processes that are particularlydamaging to soil and water resources. Theamount of sediment produced by ephemeralgully erosion, for example, can equal the

amount of sediment produced by rill andinterrill erosion in the same field (Foster,1986;Thomas et al., 1986). None of the stud-ies we reviewed include estimates ofephemeral or incised gully erosion. Waterbudgets and water supplies will be affected inimportant ways if more annual precipitation isallocated to surface runoff through increasesin precipitation intensity.

Alternatively, an increase in the number ofwet days—all other factors being equal—willcause less of an increase in erosion and runoffcompared to an increase in precipitation inten-sity, but will likely increase the importance ofsubsurface flow in delivering soluble nonpoint-source pollutants to surface water and ground-water. Hatfield et al. 1998, for example, foundthat delivery of nitrate increased linearly withcumulative weekly precipitation in a tile-drained watershed in Iowa.

Geographic, seasonal, and temporal variability ineffects will be large. All of the studies wereviewed suggest that effects of a changedprecipitation regime will vary a great deal inspace and time. Nearing 2001, for example,estimated changes in erosivity ranging from–65.1 percent to +802.8 percent among dif-ferent geographic regions in the UnitedStates. In large part, the variability in effectsreflects the large geographic, seasonal, andtemporal variability in precipitation patternsdescribed earlier. But other factors, also high-ly variable,will influence the effect of changesin precipitation patterns.

Different landscapes vary greatly in theirvulnerability to soil erosion and runoff. Soiltype, slope, and cover all have profoundeffects—both positive and negative—on theeffect of a change in precipitation regime on soilerosion, surface runoff, and other environmen-tal consequences. Spatial and temporal variabil-ity in those factors magnifies the spatial andtemporal variability in precipitation patterns.

In addition, row-crop production systemsare especially vulnerable to rainfall events thatoccur at particular times of the year. Thosetimes are (1) when the soil is most exposedbecause crops are not present or crop residuesare minimal and (2) when potential pollutantsin the soil system are at high levels and cropsare not actively growing. For example,Kramer 2001 documented the seasonalityand year-to-year variability of sediment yieldfrom a small watershed in southwestern Iowafrom 1975 to 1991. In this study, there was adistinct tendency for maximum soil erosionand runoff to occur in the early plant growth


period (spring). Thurman et al. 1991 docu-mented the “spring flush” phenomenon ofherbicide concentrations and frequency ofdetections of herbicides in the post-plantingsampling dates of surface water in the mid-western United States. Squillace andEngberg 1998 hypothesized that surfacewater containing those elevated concentra-tions of herbicides may contribute significant-ly to alluvial aquifer contamination as a resultof the reverse groundwater flow gradient (i.e.surface water from the stream flowing into thealluvial aquifer). A precipitation increase inIowa after nitrogen has been applied in thespring will result in more nitrogen in streamsthan if a similar increase in precipitation occursin August (Schuman et al., 1973).

If the seasonal distribution of precipitationchanges so that more precipitation occursduring vulnerable periods, then the effects onsoil degradation and water pollution will bemagnified. Groisman et al. 2001 found that thelargest linear trends in precipitation in theUnited States occurred in spring and autumn.The spring season is particularly vulnerablebecause of the lack of crop residues and/or agrowing crop and the fact that fertilizers and pes-ticides have been applied to agricultural fields.

In sum, a change in precipitation regimealso produces a change in the level of risk towhich agricultural land is exposed. In gener-al, a regime with greater annual precipita-tion—particularly if storm intensity increasesmore than storm frequency—heightens therisk of soil erosion, runoff, and related envi-ronmental and ecological damages. The actu-al damages that occur because of the new,more risky baseline conditions are highlydependent on additional factors.

Interactions with other components of climatechange could mitigate or exacerbate the effects ofchanges in precipitation patterns. The effect of achange in precipitation on soil erosion andrunoff and the consequent effects on theenvironment can be reduced or intensified byinteractions with other factors. Biomass is aparticularly influential factor. The increasedrisk to soil erosion and water pollution from anincrease in annual precipitation could be miti-gated by the increase in biomass expected froman increase in carbon dioxide concentrationsin the atmosphere. Similarly, the reduced riskto soil erosion and water pollution from adecrease in annual precipitation could benegated by a reduction in biomass caused by areduction in soil moisture.

Pruski and Nearing (in press) evaluated

these interactions at eight U.S. locations.They found that biomass production, asaffected by soil moisture, carbon dioxide con-centration, temperature, and solar radiation,was a key factor that sometimes overshad-owed the direct effects of rainfall increases ordecreases on soil erosion and runoff.Different types of changes occurring in dif-ferent periods of the year also complicatedsystem responses. Overall, they found thatthese interactions tended to be more impor-tant when precipitation decreased than whenprecipitation increased. Significant increases inprecipitation, they found, could be expected tolead to increased soil erosion. Decreases in pre-cipitation could produce reductions or increas-es in soil erosion, depending on the interactionof plant biomass with erosion and runoff.

Feedback mechanisms could magnify the long-term effects of a change in precipitation regime.Over time, the soil’s capacity to resist theeffects of extreme or, in some cases, meanevents can be reduced. In some cases, this lossin resilience or resistance can happen sudden-ly in response to a particularly damaging event.Runoff from an infrequent but large-magni-tude event, for example, can destabilize gullysidewalls, increasing the soil’s vulnerability tofuture, much smaller runoff events (Toy et al.,2002). An extreme erosion episode thatremoves a large portion of the organic matter-rich topsoil likely would decrease the ratewater infiltrates into the damaged soil, leadingto accelerated surface runoff, less biomass pro-duction, less soil cover and greater risk of ero-sion in future events of a smaller magnitude.

Such losses in resistance or resilience leavesoils and landscapes more vulnerable to sub-sequent events. In the worst case, this canlead to a downward spiral characterized byaccelerated degradation of soil and relatedwater resources (Seybold et al., 1999).Degradation of soil, stream channels, andother watershed features can lead to a similardownward spiral in watershed functions.

CONSERVATION IMPLICATIONSThe key findings from our review of (1) sim-ulated and observed changes in precipitationregime and (2) quantitative estimates of theeffect of changes in precipitation on soil ero-sion and runoff formed the factual basis fordiscussion and debate among our expertpanel and staff. We also relied on the profes-sional judgment of the panel and SWCS staffto answer our primary question: Are theeffects of changing precipitation regimes onsoil erosion and runoff from cropland likelyto be large enough to warrant changes in U.S.conservation policy or practice?

Our answer was, emphatically, yes.Conservationists should be concerned. Thatanswer led to further discussion about howconservationists should respond to the riskposed by climate change in the form of analtered precipitation regime.

Conservationists Should Be ConcernedConservationists should be seriously con-cerned about the implications of climatechange—as expressed by changes in precipi-tation patterns—for the conservation of soiland water resources in the United States. Themagnitude and extent of increased rates ofsoil erosion and runoff that could occurunder simulated future precipitation regimesare large. More importantly, analyses of theclimate record in the United States show thatchanges in precipitation patterns are occur-ring now. In fact, the magnitude of observedtrends in precipitation and the bias towardmore extreme precipitation events are, in somecases, larger than simulated by global climatechange models, particularly since 1970. Thesoil degradation, water pollution, and otherenvironmental effects noted in Figure 1 thatcould occur as a result of the more pronouncedincreases in precipitation amounts and intensi-ties would be severe enough that conservationpolicies and practices would need to changeand adapt to the increased risk of soil erosionand surface runoff.

Pruski and Nearing 2002, for example,quantified the relationship between a 1 percentchange in precipitation and the resultingchange in soil erosion and runoff, dependingon whether precipitation increased in frequen-cy, intensity, or a combination of both.Extrapolating those relationships to thechanges in precipitation observed over the pastcentury suggests increases in soil erosion rang-


Table 4. Potential effects on soil erosion and runoff from cropland of observed changes inprecipitation.

Increase in Mean Annual Precipitation

5% 10% 20% 40%

Change in ErosionIncrease only frequency of precipitation 4% 9% 17% 34%Increase only intensity of precipitation 12% 24% 48% 95%Increase frequency and intensity equally 8% 17% 33% 66%Change in RunoffIncrease only frequency of precipitation 6% 13% 26% 51%Increase only intensity of precipitation 13% 25% 50% 100%Increase frequency and intensity equally 10% 20% 39% 79%

Source: Derived from Pruski and Nearing 2002.

ing from 4 percent to 95 percent and increases inrunoff from 6 percent to 100 percent couldalready be evident in some locations (Table 4).

Unless additional protective measures aretaken, such increases in soil erosion on crop-land—if widespread—could reverse much ofthe progress that has been made in reducing soildegradation on cropland in the United States.

Changes of this magnitude in soil erosionand runoff likely would have profound effectson water resources, although published esti-mates are few. Nicks 1993 estimated that a 16percent average annual increase in precipita-tion,using mean monthly precipitation quan-tities from global climate change model esti-mates, would produce a 128 percent or 103percent increase in sediment yield and a 70percent or 73 percent increase in water yieldwithin an Oklahoma watershed dependingon what method was used to estimate mont-ly precipitation trends . Favis-Mortlock andGuerra 1999 used mean monthly precipita-tion outputs from three global climate mod-els for 2050 to estimate changes in sedimentyield in Sorriso, Brazil. They reported anincrease in sediment yield of 27 percentunder the first model in which change inmean monthly precipitation ranged from –10percent to 38 percent, a decrease in sedimentyield of 9 percent under the second model inwhich change in mean monthly precipitationranged from –20 percent to 4 percent, and anincrease in sediment yield of 55 percentunder the third model in which change inmean monthly precipitation ranged from –5percent to 95 percent. Transport of solublepollutants, such as nitrates, some pesticides,and salts, could be accelerated by changes inprecipitation patterns that are significantlysmaller than those needed to induce acceler-ated erosion. Hatfield et al. 1998, for exam-ple, reported that monitoring of subsurfacedrainage in an Iowa watershed indicates that67 percent of the nitrate loadings could beaccounted for by considering the effects ofprecipitation patterns in the watershed.

The potential for climate change—asexpressed in changed precipitation regimes—to increase the risk of soil erosion, surfacerunoff, and related environmental conse-quences is clear. The actual damage thatwould result from such a change is unclear.That uncertainty is largely the result of vari-ability in precipitation patterns and variabili-ty in the landscape features and managementthat ultimately determine the outcome of achange in a precipitation regime.

Regional, seasonal, and temporal variabili-ty in precipitation is large in both simulatedclimate regimes and the existing climaterecord. Different landscapes vary greatly intheir vulnerability to soil erosion and runoff.Timing of biomass production and harvest,tillage operations, and applications of nutri-ents and pesticides combine to create greatervulnerabilities to soil erosion and runoff dur-ing certain seasons. The effect of a particularstorm event depends on the moisture contentof the soil before the storm starts and, therefore,on the amounts and intensities of previousstorm events. These interactions between pre-cipitation, landscape, and management meanthe actual outcomes from any particular changein precipitation regime will be complex.

In sum, a change in precipitation regimealso produces a change in the level of risk towhich agricultural land is exposed. In gener-al, a regime with greater annual precipita-tion—particularly if increased storm intensitychanges more than storm frequency—heightens the risk of soil erosion, runoff, andrelated environmental and ecological dam-ages. In general, the risk increases at a greaterrate than precipitation amount or intensityincreases. Whether that new,more risky base-line condition translates into greater soildegradation, pollution of surface water, pollu-tion of groundwater or a combination of allthree outcomes is highly dependent on inter-actions among precipitation patterns, land-scape features, and management.

We focused our work on the effects ofincreased precipitation amounts and intensi-ties on soil erosion and runoff from cropland.It is important to note, however, that increas-es in the amount and intensity of precipita-tion are not the only—and in some casesprobably not the most important—risksposed by climate change. Decreases in pre-cipitation, for example, can accelerate erosionas much as increases in precipitation becauseof reduced biomass production. Pruski and

Nearing (in press) reported that soil erosionincreased in about half of the cases in whicha decrease in total precipitation was simulat-ed. More importantly, change in the season-al distribution of precipitation, coupled withincreased minimum temperatures, could haveprofound effects on water budgets in thewestern United States. Those effects could bemore important than effects on erosion,runoff, and related environmental endpoints.Similarly, a change in precipitation regimes inarid areas could lead to major shifts in plantcommunities on rangeland. Such shifts couldhave profound effects on the environmentand on the sustainability of agricultural pro-duction in those regions. We did not attemptto assess the magnitude of such changes, butsuch an assessment would likely increase thereasons conservationists should be concernedabout climate change.

Work to identify the most effective ways inwhich soil and water conservation policiesand practices can be adapted to a changingprecipitation regime should, therefore, beundertaken with some urgency.

How Should ConservationistsRespond?The primary objective of this project was toanswer one question: Are the effects ofchanging precipitation regimes on soil ero-sion and runoff from cropland likely to belarge enough to warrant changes in U.S. con-servation policy or practice? Our answer was,emphatically, yes. Conservationists should beconcerned. In the process of coming to thatconclusion, we developed ideas regarding themost promising responses to the risk posed byan altered precipitation regime. Three partic-ularly promising responses include:

1. Immediately update climatic parametersin critical conservation planning tools.

2.Undertake targeted investigations to firm upestimates of the damage that would likely occurunder simulated or observed climate regimes.


Regional Thresholds for Heavy, Very Heavy, and Extreme Precipitation Events

In the table above, for example, the nationwideaverage rainfall for an event defined as heavy (95thpercentile) is 30 mm and ranges from 20 mm to 45mm among the four regions analyzed. The range inabsolute amounts of rainfall is larger for eventsdefined as heavy (99th percentile) or extreme(99.9th percentile).

The same pattern occurs if heavy precipitationevents are defined in terms of duration and returnperiod. Kunkel et al. 1999a, for example, reportedthat thresholds for 7-day, 1-year recurrence rainfallevents ranged from less than 25 mm in the desertregions of the southwestern United States to morethan 150 mm along the coasts of the Gulf ofMexico, southern Atlantic Ocean, and northernPacific Ocean. The thresholds for 3-day, 1-year

recurrence events were about 80 percent of thosefor 7-day, 1-year events; thresholds for 7-day, 5-yearrecurrence events were about 150 percent of thosefor 7-day, 1-year events.

The absolute intensity and duration of rainfallevents are the most important factor in determiningwhether an increasing trend in such events willresult in increased erosion, runoff, and related envi-ronmental damage. A large upward trend in heavyprecipitation events could cause little damage if theevents remain very rare or if the absolute amount ofrainfall during the events is small. On the otherhand, even very rare events can cause serious andlong-lasting damage if rainfall intensities and dura-tion are large.

The absolute amount and intensity of rainfall that occurs in events defined as heavy or extreme vary acrossthe United States. In general, smaller absolute amounts of precipitation occur during heavy, very heavy, orextreme events in regions with less annual rainfall.

Amount of 24-hour Precipitation by Region (mm)

Type of Missouri SoutheasternPrecipitation Southwest River Midwest Coastal

Nationwide

Event Basin regionAverage

Heavy 20 20 35 45 30(above 95th

percentile)

Very heavy 34 37 59 82 52(above 99th

percentile)

Extreme 56 64 99 151 89(above 99.9th

percentile)

Source: P.Y. Groisman, U.S. National Climatic Data Center.

3. Evaluate the need to build the risk ofdamage from severe storms into the conserva-tion planning process used to determine if rec-ommended conservation practices and systemsadequately protect soil and water resources.

Update Climatic ParametersA step that could and should be taken quick-ly is to adjust the critical parameters in con-servation planning tools to reflect the most cur-rent climate data. Key climate components ofthe Revised Universal Soil Loss Equation(RUSLE), for example, are currently derivedfrom climatic data for the period 1936 to 1957over much of the contiguous United States.NRCS already is working on a new version ofRUSLE (RUSLE2) that will feature updatedalgorithms for computing soil loss on cropland.

New precipitation and temperature datacovering the period 1971-1999 are nowbeing prepared for RUSLE2 implementation,including updated R-factors for all parts ofthe continental United States. InterpolatedR-factors, as well as other required climate ele-ments,will be derived for counties, agriculturalregions, or other homogeneous areas using thePRISM spatial climate modeling system fromOregon State University. NRCS managersexpect to have this new climate informationintegrated into RUSLE2 sometime in 2003.

Precipitation frequency studies are beingupdated by the NOAA-NWSHydrometerological Design Studies Centerwith support from NRCS in the U.S.Department of Agriculture, the U.S. ArmyCorps of Engineers, and others. The originalprecipitation frequency studies were devel-oped for the eastern United States and pub-lished as Technical Paper-40 (Hershfield,1961).Technical Paper-40 was superseded inthe West by NOAA Atlas 2 (Miller et al.,1973) and Technical Memorandum NWSHydro 35 (Frederick et al., 1977). Variousother documents extended the precipitationfrequency studies to specific geographic areasof the United States. Data used in these doc-uments are what were available immediatelyprior to publication.

An updated precipitation frequency studyfor the southwestern United States (Nevada,Arizona, Utah, New Mexico, and southeast-ern California) was scheduled for release in2002. An updated precipitation frequencystudy for the Ohio River Valley region is tobe released in 2003. A nationwide update ofprecipitation frequency is planned to begin in2003 or 2004.

Improve Damage EstimatesThe factors determining the actual outcomesfrom a change in precipitation regime on soilerosion, runoff, and related environmentalresources is complex, as noted above. Targetedinvestigations should be undertaken to reducethe uncertainties in damage estimates that arisefrom those complexities. Such investigationswould provide a much firmer foundation onwhich to evaluate the changes in U.S. conser-vation policies and practices that are warranted,feasible, and cost-effective.

Our expert panel members discussed anumber of such investigations that wouldhelp inform policymakers and program man-agers. They included:

• Replicating studies, such as those cur-rently in the literature that specificallyevaluated the effect of precipitation pat-terns on soil erosion and runoff, at morelocations and land uses to develop betterestimates of the variability induced bydifferences in geography, hydrology, sea-

son, and agricultural production systems.• Utilize the NRCS National Resources

Inventory framework to update nationaland regional estimates of soil erosion,runoff, and related environmental out-comes based on updated climate param-eters in USLE or RUSLE.

• Utilize the NRCS National ResourcesInventory framework to conduct sensitiv-ity analyses and simulate the effect of sim-ulated climate scenarios on national andregional estimates of soil erosion, runoff,and related environmental outcomes.

• Complete targeted evaluations and stud-ies to better understand the causes of,future trends in, and expected magni-tudes of changes in precipitation regime,particularly those portions of the regimeof most concern for soil erosion, runoff,and related environmental effects.


Conservation Planning as Risk ManagementThere is strong agreement among modeledclimate projections that precipitation intensi-ty will increase at a rate greater than the rateof increases in mean or total annual precipita-tion. This trend toward more precipitationoccurring in more extreme events and short-er return periods for heavy precipitationevents is evident in the existing U.S. climaterecord, particularly since 1970.

The erosivity of heavy precipitation eventscan be very large because erosivity is stronglycorrelated to the product of total rainstormenergy and maximum 30-minute rainfallintensity. Even small increases in mean pre-cipitation, therefore, could result in muchgreater damage if most of that precipitationincrease comes in more intense storms. Stormintensity also increases the risk of ephemeralgully erosion, incised gully erosion, and streamchannel erosion. These forms of erosion cancause severe and lasting damage to soil and waterresources and require more intensive and oftencostly conservation treatments. Heavy precipita-tion events also can increase the vulnerability oflandscapes and watersheds to future, less intense,and more frequent precipitation events.

In current practice, conservation systemsapplied to agricultural land are usuallydesigned, planned, and implemented toaddress soil erosion, runoff, and related effectsunder estimates of expected average annualclimate regimes. Several critiques of currentpractice (Larson et al., 1997; Kramer, 2001)suggest that conservation systems shouldinstead be designed to resist the effects ofextreme events of some designated returnperiod. This approach is currently used whendesigning engineering structures to resist 25-year, 24-hour storm events, for example.

Proponents who desire a more risk-basedapproach to conservation planning point outthat most damage to soil resources occursduring infrequent but severe events. Larsonet al. 1997, for example, reported that 60 per-cent of the total soil erosion that occurred intheir Minnesota study location over 10 yearsoccurred during one cropping season.Precipitation amounts and intensities duringthat cropping season increased the erosivity ofthe study area by almost fourfold over thelong-term average. Similar results werereported by Burwell and Kramer 1983 inMissouri; Hjemfelt et al. 1986; Edwards andOwens 1991; Zuzel et al. 1993 in Oregon;Moldenhauer and Wischmeier 1960; and

Barnett and Hendrickson 1960 in Georgia.Similar arguments can be made for the dis-proportionate effect of extreme events onwater resources, particularly with respect tosediment delivery, channel stability, and otherwatershed processes.

The adequacy of current conservationplanning approaches based on annual averageprecipitation should be reevaluated in light ofthe evidence that future—or current—cli-mate regimes are characterized by increasedfrequency of extreme events. Such a changein conservation planning would require:

• Identifying and quantifying thresholds ofsoil erosion, runoff, or transport of pol-lutants above which damage at somespecified frequency is unacceptable.

• Identifying critical times during the yearwhen the nature and timing of agricul-tural production practices increase therisk of damage.

• Determining the probability of occur-rence of storms of sufficient intensityand timing to cause damage that exceedsthreshold levels.

• Designing conservation systems to beprotective at vulnerable times and dur-ing particular events.

Such an approach is routinely used whenengineering structural solutions to soil ero-sion and runoff, such as terraces or grassedwaterways. Such approaches are not routine-ly used when planning conservation systemsthat rely more on annual conservation prac-tices, such as conservation tillage or nutrientmanagement. The environmental perform-ance of agricultural production systemsdepends heavily on how well such annualpractices perform.

Research and technology developmentwill be required to provide the necessaryunderstanding and tools to make a risk-basedapproach to conservation a reality at the fieldlevel. Current field-based models may needto be updated or replaced with alternativesthat allow planners to incorporate the proba-bility of extreme events into their recom-mendations for reducing soil erosion, runoff,and water pollution. Similarly, the perform-ance of existing and new conservation sys-tems would need to be assessed regardingtheir performance during extreme events.Such an assessment might, for example,enhance the importance of conservation sys-tems that encourage continuous soil cover;utilize filter strips, riparian buffers, or vegeta-tive barriers; or minimize the concentration

of nutrients, pesticides, salts, or other potentialpollutants during vulnerable seasons.

The first step in that effort should be aseries of initial risk-based assessments of therelative contribution of severe storms to soildegradation and water pollution. Theseassessments should be targeted at quantifyinghow well key conservation systems protectkey environmental endpoints under severestorms. The results of these assessmentsshould be used to determine (1) if a shift to amore risk-based conservation planningapproach is warranted based on the risk ofenvironmental damage in the absence of sucha shift, (2) to what extent such a risk-basedconservation planning approach wouldchange the conservation practices and systemsrecommended for implementation on U.S.farms and ranches, (3) the steps needed toimplement such an approach, and (4) a pre-liminary estimate of the cost of implementingsuch an approach.

The initial assessments should be under-taken quickly. The updated and improved cli-mate databases that are being assembledshould facilitate such evaluations. The basicmodeling and prediction capability neededfor such assessments is already available incontinuous-simulation, storm-based models.Resources should be made available to con-duct the assessments. Those resources willleverage the investments already made inassembling data and improving models.Policymakers and program managers needanswers soon so they can begin to make theinvestment decisions and changes in opera-tional policy needed to ensure that U.S. soiland water resources are protected under achanging climate.


REFERENCESArnold, J.G., J.R. Williams, A.D. Nicks, and N.B.

Sammons. 1990. SWRRB: A Basin ScaleSimulation Model for Soil and WaterManagement. Texas A&M University Press,College Station. 142 pp.

Barnett, A.P. and B.H. Hendrickson. 1960. Erosionon Piedmont soils. Soil Conservation 26:31-34.

Boardman, J. and D.T. Favis-Mortlock. 1993.Climate change and soil erosion in Britain.TheGeographical Journal 159(2):179-183.

Brown, R.D. and R.O. Braaten. 1998. Spatial andtemporal variability of Canadian monthly snowdepths,1946-1995.Atmosphere-Ocean 36:37-45.

Burwell, R.E. and L.A. Kramer. 1983. Long-termannual runoff and soil loss from conventionaland conservation tillage of corn. Journal of Soiland Water Conservation 38(3):315-319.

Cayan, D.R., S. Kammerdiener, M.D. Dettinger,J.M. Caprio, and D.H. Peterson. 2001. Changesin the onset of spring in the western UnitedStates. Bulletin of the American MeteorologicalSociety 82(3):399-415.

Changnon, S.A. 1998. Comments on “Seculartrends of precipitation amount, frequency, andintensity in the United States.” Bulletin of theAmerican Meteorological Society 79:2550-2552.

Easterling, D.R., T.R. Karl, E.H. Mason, P.Y.Hughes, D.R. Bowman, R.C. Daniels, and T.A.Boden. 1996. United States HistoricalClimatology Network (U.S. HCN). Monthlytemperature and precipitation data. ESD Publ.4500, ORNL/CDIAC/-87, NDP-019/R3 83pp. plus appendixes. [Available from CarbonDioxide Information Analysis Center, OakRidge National Laboratory, Oak Ridge, TN37831-6335].

Easterling, D.R., T.R. Karl, K.P. Gallo, D.A.Robinson, K.E.Trenberth, and A. Dai. 2000a.Observed climate variability and change of rel-evance to the biosphere. Journal ofGeophysical Research 105(D15):101-114.

Easterling, D.R., G.A. Meehl, C. Parmesan, S.A.Changnon,T.R. Karl, and L.O. Mearns. 2000b.Climate extremes: Observations, modeling, andimpacts. Science 289(5487): 2068-2074.

Easterling, D.R., J.L. Evans, P.Ya. Groisman, T.R.Karl, K.E. Kunkel, and P. Ambenje. 2000c.Observed variability and trends in extreme cli-mate events: A brief review. Bulletin of theAmerican Meteorological Society 81(3):417-425.

Edwards,W.C. and L.B. Owens. 1991. Large stormeffects on total soil erosion. Journal of Soil andWater Conservation 46(1):75-78.

Favis-Mortlock, D.T., R. Evans, J. Boardman, and T.M. Harris. 1991. Climate change, winter wheatyield, and soil erosion on the English SouthDowns.Agricultural Systems 37:415-433.

Favis-Mortlock, D.T. and M.R. Savabi. 1996. Shiftsin rates and spatial distribution of soil erosionand deposition under climate change. Pp. 529-560. In: M.G. Anderson and S.M. Brooks (eds.).

Advances in Hillslope Processes.Wiley,Chichester.Favis-Mortlock, D.T. and A.J.T. Guerra. 1999.The

implications of general circulation model esti-mates of rainfall for future erosion:A case studyfrom Brazil. Catena 37(3-4):329-354.

Foster, G.R. 1986. Understanding ephemeral gullyerosion. Pp. 90-125. In: Soil Conservation:Assessing the National Resources Inventory.National Academy Press, Washington, DC.

Foster, G.R., L.J. Lane, and M.A. Nearing. 1989.Erosion component. Pp. 10.1-10.12. In:USDA-Water Erosion Prediction Project,Hillslope Profile Model Documentation.NSERL Report No. 2. U. S. Department ofAgriculture, Agricultural Research Service,National Soil Erosion Research Laboratory,West Lafayette, Indiana.

Foster, G.R., R.A.Young, M.J. M. Romkens, andC.A. Onstad. 1985. Processes of erosion bywater. Pp. 137-162. In: R. F. Follett and B. A.Stewart (eds). Soil Erosion and CropProductivity. American Society of Agronomy,Crop Science of America and Soil Science ofAmerica, Madison,Wisconsin.

Frederick, R.H., V.A. Myers, and E.P. Auciello.1977. Five- to 60-minute precipitation fre-quency for the eastern and central UnitedStates. Technical Memorandum NWSHYDRO-35. U.S. Department of Commerce,Silver Spring, Maryland.

Frei,A.,D.A.Robinson,and M.G.Hughes.1999.NorthAmerican snow extent, 1910-1994. InternationalJournal of Climatology 19:1517-1534.

Green,W.H., and G.A. Ampt. 1911. Studies in soilphysics I: The flow of air and water throughsoils. Journal of Agricultural Sciences 4:1-24.

Groisman, P.Y. and D.R. Easterling. 1994.Variabilityand trends of precipitation and snowfall overthe United States and Canada. Journal ofClimate 7:184-205.

Groisman, P.Y., T.R. Karl, and R.W. Knight. 1994.Observed impact of snow cover on the heat bal-ance and the rise of continental spring tempera-tures. Science 263:198-200.

Groisman, P.Y., D.R. Easterling, R.G. Quayle, P.Ambenje, and R.W. Knight. 1999a. Trends inprecipitation and snow cover in the UnitedStates. Pp.89-92. In: D.B.Adams (ed.). PotentialConsequences of Climate Variability andChange to Water Resources of the UnitedStates.American Water Resources Association.

Groisman, P.Y., T.R. Karl, D.R. Easterling, R.W.Knight, P.F. Jamason, K.J. Hennessy, R. Suppiah,C.M. Page, J. Wibig, K. Fortuniak, V.N.Razuvaev, A. Douglas, E. Forland, and P.M.Zhai. 1999b. Changes in the probability ofheavy precipitation: Important indicators of cli-matic change. Climatic Change 42:243-283.

Groisman, P.Y., R.W. Knight and T.R. Karl. 2001.Heavy precipitation and high streamflow in thecontiguous United States:Trends in the twenti-eth century. Bulletin of the AmericanMeteorological Society 82(2):219-246.

Guttman, N.B. and R.G. Quayle, 1996:A historicalperspective of U.S. Climate Divisions. Bulletinof the American Meteorological Society77:293-303.

Hansen, J., I. Fung,A. Lacis, D. Rind, S. Lebedoff, R.Ruedy, and G. Russell. 1988. Global climatechanges as forecast by Goddard Institute forSpace Studies three-dimensional model. Journalof Geophysical Research 93:9341-9364.

Hatfield, J.L., J.H. Prueger, and D.B. Jaynes. 1998.Environmental impacts of agricultural drainage inthe Midwest. Pp. 28-35. In: L.C. Brown (ed.).Drainage in the 21st Century: Food Productionand the Environment. Proceedings of the 7thannual drainage symposium, 8-10 Mar. 1998,Orlando, Florida. American Society ofAgricultural Engineers, St. Joseph, Michigan.

Hershfield, D.M. 1961. Rainfall frequency atlas ofthe United States for durations from 30 minutesto 24 hours and return periods from 1 to 100years. Technical Paper No. 40. Weather Bureau,U.S.Department of Commerce,Washington,D.C.

Hjelmfelt,A.T. Jr., L.A. Kramer, and R.G. Spooner.1986. Role of large events in average soil loss. Pp.3.1-3.9.In:Proceedings,Fourth Federal InteragencySedimentation Conference, Volume. 1. U.S.Geological Survey,Denver,Colorado.

IPCC (Intergovernmental Panel on ClimateChange), Working Group I. 2001a. ClimateChange 2001:The scientific basis.Contribution ofWorking Group I to the Third Assessment Reportof the IPCC. Cambridge University Press.

IPCC (Intergovernmental Panel on ClimateChange), Working Group I. 2001b. ClimateChange 2001: Summary for Policy Makers.Cambridge University Press.

Jasa, P.J., E.C. Dickey, and D.P. Shelton. 1986. Soilerosion from tillage and planting systems used insoybean residue: Part 2. Influences of row direc-tion. Transactions, American Society ofAgricultural Engineers 29:761-766.

Karl, T.R., and R.W. Knight. 1998. Secular trendsof precipitation amount, frequency, and intensi-ty in the USA. Bulletin of the AmericanMeteorological Society 79:231-241.

Karl, T.R., R.W. Knight, D.R. Easterling, and R.G.Quayle. 1996. Indices of climate change for theUnited States. Bulletin of the AmericanMeteorological Society 77: 279-292.

Kramer, L.A. 2001. Seasonal soil loss from a smallHEL watershed. Pp. 215-218. In: J. C. AscoughII and D. C. Flanagan (eds.). Soil ErosionResearch for the 21st Century. Proceedings,International Symposium, Honolulu, Hawaii,3-5 January 2001. American Society ofAgricultural Engineers, St. Joseph, Michigan.

Kunkel, K., K.Andsager, and D.R. Easterling. 1999a.Long-term trends in heavy precipitation eventsover the continental United States. Journal ofClimate 12:2515-2527.

Kunkel, K., R.A. Pielke Jr., and S.A. Changnon.1999b.Temporal fluctuations in weather and cli-mate extremes that cause economic and human


health impact:A review. Bulletin of the AmericanMeteorological Society 80:1077-1098.

Larson,W.E., M.J. Lindstrom, and T.E. Schumacher.1997.The role of severe storms in soil erosion:A problem needing consideration. Journal ofSoil and Water Conservation 52(2):90-95.

Lee, J.J., D.L. Phillips, and R.F. Dodson. 1996.Sensitivity of the US corn belt to climate changeand elevated CO2: II. Soil erosion and organiccarbon.Agricultural Systems 52:503-521.

Manabe, S. and R.T. Wetherald. 1987. Large-scalechanges in soil wetness induced by an increasein carbon dioxide. Journal of AtmosphericSciences 44:1211-1235.

Miller, J.F., R.H. Frederick, and R.J. Tracey. 1973.Precipitation-frequency atlas of the westernUnited States,Vol. I-XI,NOAA Atlas 2,NationalWeather Service, National Oceanic andAtmospheric Administration, U. S. Departmentof Commerce, Silver Spring, Maryland.

Mitchell, J.F., C.A.Wilson, and W.M. Cunningham.1987. On CO2 climate sensitivity and modeldependence of results. Quarterly Journal of theRoyal Meteorological Society 113:293-322.

Moldenhauer, W. C. and W. H. Wischmeier. 1960.Soil and water losses and infiltration rates on Idasilt loam as influenced by cropping systems,tillage practices and rainfall characteristics. SoilScience Society of America Journal 24:409-413.

Nearing, M.A. 2001. Potential changes in rainfallerosivity in the U.S. with climate change dur-ing the 21st century. Journal of Soil and WaterConservation 56(3):229-232.

Nicks, A.D. 1993. Modeling hydrologic impacts ofglobal change using stochastically generated cli-mate change scenarios. Pp. 25-38. In: Y.Eckstein and A. Zaporozec (eds.). Proceedingsof Industrial and Agricultural Impacts on theHydrologic Environment,Volume. 3, Impact ofEnvironmental and Climatic Change on Globaland Regional Hydrology. Water EnvironmentFederation,Alexandria,Virginia.

Nicks, A.D. and G.A. Gander. 1994. CLIGEN: Aweather generator for climate inputs to waterresource and other models. Pp. 3-94. In:Proceedings of the 5th InternationalConference on Computers in Agriculture,1994. American Society of AgriculturalEngineers, St. Joseph, Michigan.

Oldeman, L.R. 1994. The global extent of soildegradation. Pp. 99-118. In: D.J. Greenland andI. Szabolcs (eds.). Soil Resilience andSustainable Land Use. CAB International,Wallingford, U.K.

Peterson, T.C., V.S. Golubev, and P. Ya Groisman.1995. Evaporation losing its strength. Nature377:687-688.

Phillips, D.L., D. White, and B. Johnson. 1993.Implications of climate change scenarios for soilerosion potential in the USA. LandDegradation & Rehabilitation 4:61-72.

Pruski, F.F. and M.A. Nearing. 2002. Runoff and

soil-loss responses to changes in precipitation:Acomputer simulation study. Journal of Soil andWater Conservation 57(1): 7-16.

Pruski, F.F. and M.A. Nearing (in press). Climate-induced changes in erosion during the 21stcentury for eight U. S. locations. WaterResources Research.

Renard, K.G., G.R. Foster, G.A. Weesies, D.K.McCool, and D.C. Yoder, coordinators. 1997.Predicting soil erosion by water:A guide to con-servation planning with the revised universal soilloss equation (RUSLE). U.S. Department ofAgriculture,Agricultural Handbook No. 703.

Renard, K.G. and J.R. Friedmund. 1994. Usingmonthly precipitation data to estimate the R-factor in the revised USLE. Journal ofHydrology 157: 287-306.

Savabi, M.R., J.G. Arnold, and A.D. Nicks. 1993.Impact of global climate changes on hydrologyand soil erosion: A modeling approach. Pp. 3-18. In: Y. Eckstein and A. Zaporozec (eds.).Proceedings of Industrial and AgriculturalImpacts on the Hydrologic Environment,Vol. 3,Impact of Environmental and Climatic Changeon Global and Regional Hydrology. WaterEnvironment Federation,Alexandria,Virginia.

Schlesinger, M.E. and Z.C. Zhao. 1989. Season cli-matic change introduced by doubled CO2 assimulated by the OSU atmosphericGCM/mixed-layer ocean model. Journal ofClimate 2:429-495.

Schuman, G.E., R.E. Burwell, R.F. Piest, and R.G.Spomer.1973.Nitrogen losses in surface runoff fromagricultural watersheds on Missouri Valley loess.Journal of Environmental Quality 2(2): 299-302.

Seybold, C.A., J.E. Herrick, and J.J. Brejda. 1999.Soil resilience: A fundamental component ofsoil quality. Soil Science 164(4): 224-234.

Sharpley, A.N. and J.R. Williams. 1990. EPIC-Erosion Productivity Impact Calculator I.Model Documentation. U.S. Department ofAgriculture Technical Bulletin 1768.

Sharpley, A.N., M.J. Hedley, E. Sibbesen, A.Hillbricht-Ilkowska, A.A. House, and L.Ryszkowski. 1995. Phosphorus transfers fromterrestrial to aquatic systems. Chapter 11. In: H.Tiessen (ed.). SCOPE 54 Phosphorus in theglobal environment – transfers, cycles and man-agement. John Wiley & Sons, Inc., New York,New York.

Squillace, P.J. and R.A. Engberg. 1998. Surface-water quality of the Cedar River Basin, Iowa-Minnesota, with emphasis on the occurrenceand transport of herbicides, May 1984 throughNovember 1985. Water-ResourceInvestigations Report (U. S. Geological Survey)No. 88-4060.

Thomas, A.W., R. Welch, and T. R. Jordan. 1986.Quantifying concentrated flow erosion oncropland with aerial photography. Journal ofSoil and Water Conservation 41(4):249-252.

Thurman. E.M., D.A. Goolsby, M.T. Meyer, and

D.W. Kolpin. 1991. Herbicides in surface watersof the midwestern United States: The effect ofspring flush. Environmental Science andTechnology 25(10):1794-1796.

Toy,T.J., G.R. Foster, and K.G. Renard. 2002. SoilErosion: Processes, Prediction, Measurementand Control. John Wiley & Sons, Inc., NewYork, New York. Pp. 64.

U.S. Department of Agriculture. 1997. NationalResources Inventory.

U.S. Department of State. 2002. U.S. climate actionreport 2002. U.S. Government Printing Office,Washington, D.C.

U.S. NAST (National Assessment Synthesis Team).2001. Climate change impacts on the UnitedStates: The potential consequences of climatevariability and change. Foundation Report.U.S. Global Change Research Program.

Williams, J.R. 1975. Sediment-yield predictionwith the universal equation using runoff ener-gy factor. Pp. 244-252. In: Present and prospec-tive technology for predicting sediment yieldsources. Proceedings of the sediment-yieldworkshop, 28-30 November 1972, U.S.Department of Agriculture, SedimentationLaboratory, Oxford, Mississippi.ARS-S-40.

Williams, J.R., C.A. Jones, and P.T. Dyke. 1984. Amodeling approach to determining the rela-tionship between erosion and soil productivity.Transactions,ASAE 27(1):129-144.

Williams, J., M.A. Nearing, A. Nicks, E. Skidmore,C. Valentine, K. King, and R. Savabi. 1996.Using soil erosion models for global changestudies. Journal of Soil and Water Conservation51(5):381-385.

Wischmeier, W.H., and D.D. Smith. 1965.Predicting rainfall-erosion losses from croplandeast of the Rocky Mountains: Guide for selec-tion of practices for soil and water conserva-tion. U. S. Department of Agriculture,Agricultural Handbook No. 282.

Wischmeier, W.H. and D.D. Smith. 1978. Predictingrainfall erosion losses:A guide to conservation plan-ning. U.S. Department of Agriculture,AgriculturalHandbook No. 537.

Young, R.A.,A.E. Olness, C.K. Mutchler, and W.C.Moldenhauer. 1986. Chemical and physicalenrichments of sediment from cropland.Transactions, American Society of AgriculturalEngineers 29(1):165-169.

Zuzel, J.F., R.R. Allmaras, and R.N. Greenwalt.1993. Temporal distribution of runoff and soillosses at a site in northeastern Oregon. Journalof Soil and Water Conservation 48(4):373-378.

Zwiers, E.W. and V. Kharin. 1998. Changes in theextremes of the climate simulated by CCCGCM2 under C02 doubling. Journal ofClimate 11(9): 2200-2222.


APPENDIXPredicting Soil Erosion and Runoff from Cropland

The studies cited in this report used one of threebasic models to estimate the effects of a changein precipitation regime on soil erosion andrunoff from cropland. The RUSLE2 model isbriefly described here because it will supercedeRUSLE. It is provided for comparative andinformational purposes.

Revised Universal Soil Loss Equation(RUSLE)Abstract (empirical), deterministic, continuous,lumped, annual time step

RUSLE is the primary field tool used byNatural Resources Conservation Service fieldstaff in conservation planning in the UnitedStates. RUSLE is based on the Universal SoilLoss Equation (USLE)—an empirical equationdeveloped by Wischmeier and Smith 1965, 1978.USLE was first published in 1965 and laterupdated in 1978.The 1978 version is document-ed in Agricultural Handbook 537.This equationwas developed to estimate the long-term sheetand rill erosion rates on cropland and pasture inthe eastern two-thirds of the contiguous UnitedStates. Later, it was adapted for use in other areasof the world and on rangeland and forestland.

Model ComponentsThe Revised Universal Soil Loss Equation

(RUSLE) (Renard et al., 1997) is described inAgriculture Handbook No. 703. RUSLE hasthe same factors as USLE, but it expands theadaptability and utility of USLE. The RUSLEequation is:

A = R * K * LS * C * Pwhere:

A = estimated average annual soil loss, t/haR = rainfall and runoff factor,

MJ*mm*ha-1*h-1*y-1

K = soil erodibility factor, soil loss per unit ofrainfall erosivity index from bare fallowon a 9% slope 22.1 m long,t*h*MJ-1*mm-1

LS = slope length and steepness factor, theratio of soil loss from a slope of specifiedsteepness and length to that of a 9% slope22.1 m long, dimensionless

C = cropping and management factor, theratio of soil loss from a specified crop-ping sequence and tillage system to thatof a bare fallow condition,dimensionless

P = erosion control support practice factor, theratio of soil loss from a slope using a soilerosion control support practice (i.e. con-touring) to the same slope conditions tilledup and down the slope, dimensionless

Factors Impacted by a Change in Rainfall IntensityFactors of RUSLE that may need to be modi-

fied or adjusted due to a change in rainfall inten-sity for a given geographic location could includethe R-factor, the K-factor, the C-factor, and theP-factor.

The R-factor. Raindrops striking the surfaceof the soil have kinetic energy because of theirmass and velocity. Observations by Wischmeierand Smith determined that the product of totalstorm energy (E) and the maximum 30-minuterainfall intensity (I30) was proportional to thesoil erosion rate. The EI30 for an individualstorm is the R-factor for that storm.The annu-al average of the long-term summed EI30 valuesof rainstorms at a locale equals the R-factor forthat specific locale. Precipitation occurring assnow is not included. In the western UnitedStates, all rainstorms are included to calculatethe R-factor. In the eastern United States, onlystorms exceeding 13 mm are used in calculatingthe R-factor unless the maximum 15-minuteintensity exceeds 25 mm h-1.

A sufficient increase in rainfall intensity willincrease the I30 value and probably the E value aswell.An increase in the I30 value will increase theR-factor proportionally.The amount of rain pro-duced by a given storm will affect the E value.Increasing the number of wet days will increasethe average R-factor, assuming rainfall quantityor intensity exceeds the thresholds listed above.

The K-factor. The soil erodibility factor is ameasured (or estimated) value of the averageannual amount of soil loss per unit of R-factor.The conditions required to measure a K-valueinclude a bare fallow plot tilled up and downthe hill at the same time of year as continuousrow-cropped fields with subsequent secondarytillage to control weeds and eliminate serious crustformation. The K-factor is a lumped parameterrepresenting the average annual response of the soiland soil profile to a large number of erosion andhydrologic processes. A change in the rainfallintensity could also affect the long-term responseof the soil and soil profile,thus changing the K-fac-tor. It also is anticipated that changes in the sea-sonality of rainfall or the number of wet days couldalso affect the soil erodibility factor as well.

The C-factor. The cropping and managementfactor is expressed as the ratio of the soil lossexpected from a specified crop/croprotation/tillage/residue system relative to the soilloss expected from a clean-tilled, continuous fal-low condition tilled up and down the slope.Thissoil loss ratio (SLR) is calculated on a 15-daytime step. During a time step, it is assumed noappreciable changes occur in the erosive condi-

tion of the soil.This time step can be segment-ed into shortened time intervals if a sudden changeoccurs that affects the soil/vegetation/residue sys-tem (e.g. tillage operation).The SLR is calculat-ed as a function of five subfactors.These subfac-tors are prior land use, canopy cover, surfacecover, surface roughness, and soil moisture. Eachof these subfactors can affect the soil/vegeta-tion/residue system that influences the SLR. Achange in rainfall intensity may directly influencethe surface roughness and soil moisture subfac-tors. Surface roughness will tend to decrease asthe EI30 value increases and lead to more runoff.As antecedent moisture increases, surface runoffwill tend to increase.

The P-factor. The soil conservation supportfactor is affected by practices that modify thesurface runoff pattern, direction or velocityand/or affect total runoff volume and/or peakrate of runoff. Previous field studies(Moldenhauer and Wischmeier, 1960; Jasa et al.,1986) indicate that contouring is less effectivefor large storms compared to smaller storms dueto a combination of total runoff volume andrainfall intensity. Renard et al. 1997, p. 191, sug-gested that the P-factor rating given a support prac-tice be based upon the storm erosivity value of aspecific probability of recurring rather than thelong-term rainfall erosivity factor.Therefore, undera climate change scenario that includes increasedrainstorm intensity or volume of runoff, the effec-tiveness of conservation support practices woulddecrease, causing an increase of soil erosion.

Revised Universal Soil Loss Equation, ver-sion 2 (RUSLE2)Abstract (empirical) and physically-based, deter-ministic, continuous, daily time step

RUSLE2 is the new field tool to be used bythe Natural Resources Conservation Service in2003. It is a Windows-based model that will beplaced in field offices for use by field personnelfor conservation planning. This soil loss equationestimates long-term sheet and rill erosion, butcalculates factors in a daily time step. Because ofthe daily time step, the stand-alone R, K, LS, C,and P factors that were the basis for the USLEand RUSLE are no longer viable as more physi-cally-based science is used.

Factors Impacted by a Change in Rainfall IntensityThe erosivity factor (R) consists of monthly

values that are disaggregated into daily values.The erodibility factor (K) varies seasonally, basedon daily temperature and precipitation. Thecover and management factor (C) uses subfac-tors similar to those in RUSLE; those subfactors


are affected by daily temperature and precipita-tion. The support practice factor (P) is upgrad-ed from RUSLE. For example, the 10-year, 24-hour storm values are used to compute runoffand the reduction in erosivity by ponding. It isused indirectly to compute factors for contour-ing, critical slope length, sediment transportcapacity, deposition, and detached soil particlesize information.

Modified Universal Soil Loss Equation(MUSLE)Abstract (empirical), deterministic, event-based,lumped

The MUSLE equation was developed by J.R.Williams 1975 as a method to estimate sedi-ment delivered from small watersheds for indi-vidual storms.This equation replaces the R-fac-tor of the USLE (or RUSLE) with a total runoffand peak runoff rate term for an individualstorm event.The MUSLE is written as:

A = 11.8 * (Qsurf * qpeak)0.56 * K * LS * C * Pwhere:

A = tons of sediment delivered to the nearest permanent channel, t

Qsurf = surface runoff depth, mmqpeak = peak runoff rate, m3 s-1

The constant 11.8 is a unit conversion factor.All other factors in the equation are defined thesame as in the USLE.

Factors Impacted by a Change in Rainfall IntensityBecause MUSLE estimates runoff and sedi-

ment delivery produced by individual storms, achange in rainfall quantity will be accountedfor in the equation with the Qsurf and qpeak fac-tors. The Soil Conservation Service CurveNumber Method is used to estimate Qsurf and isnot sensitive to rainfall intensity. Like theUSLE, however, the K-, C-, and P-factors alsomay be affected by a change in rainfall, requir-ing their adjustment.

Water Erosion Prediction Project (WEPP)Physically-based, deterministic, continuous orevent-based, distributed, daily time step

The Water Erosion Prediction Project wasinitiated in 1986 (Foster et al., 1989). It incorpo-rates the latest technology and soil erosionresearch information.WEPP can be used to pre-dict sheet and rill erosion on hillslopes and sedi-ment delivery from fields and small watersheds.The six major components of WEPP are climategeneration, hydrology, plant growth, soils, irriga-tion, and soil erosion.

The climate generator model (CLIGEN)(Nicks and Gander, 1994) can use existingweather records to simulate physical processes,or it can generate site-specific rainfall amount,duration, maximum intensity, time to peakintensity, maximum and minimum temperature,solar radiation, wind speed, wind direction, anddew point temperature. If weather records areprovided, WEPP can use breakpoint precipita-tion data or daily rainfall amounts. If daily rain-fall amounts are entered, WEPP will disaggre-gate the rainfall into a simple single-peak stormpattern. The rainfall amount and intensity dataare used to estimate the amount of infiltrationand runoff for each storm.

The hydrology component includes infiltra-tion, runoff, evapotranspiration, interception,lateral flow, and deep percolation.The water bal-ance is computed daily. Infiltration is modeledusing the Green-Ampt equation (Green andAmpt, 1911). Leaf area index calculated by thecrop growth model is used to estimate transpira-tion losses. Soil water is routed through the soillayers, eventually to percolate through the soilprofile, to enter tile drains, or to be extracted byplants. Currently, the water quality subroutinesare not incorporated in WEPP to simulate nutri-ent and pesticide movement and transformations.

The plant growth component is patternedafter the Erosion Productivity Impact Calculator(EPIC) crop growth model (Williams et al.,1984).The complete plant growth cycle, from seed tosenescence, and residue decomposition is simulat-ed.Annual and perennial plants can be simulated.

The soils component is the SOILS-5 data-base.This database includes the soil parametersnecessary for calculating hydrologic and erosioninputs.The model changes these parameters on thedaily time step to account for temporal changes ofthese properties due to tillage operations, freezingand thawing, compaction, and precipitation.

The irrigation component can simulate sprin-kler or furrow irrigation.The schedule of irriga-tion is user-defined by one of three methods.Sprinkler irrigation is modeled as a rainfall event ofeven intensity.

The erosion component uses the steady statesediment continuity equation as the basis forcomputing erosion. Soil detachment and inter-rill sediment delivery to the rills are modeled asa function of rainfall intensity and slope. Interrilldetachment and transport and rill erosion aremodeled separately. Soil detachment in rillsoccurs if the hydraulic shear stress of the wateris greater than the critical flow shear stress of thesoil and the sediment load of the flow is less thanthe flow’s transport capacity. Deposition occurs

if the sediment load exceeds the capacity of theflow to carry it. Sediment is characterized byparticle size based upon the soil profile fromwhich it eroded. The preferential sorting andtransport of soil particles that are smaller andlighter is the basis for enrichment ratios forclays, organic matter, and nutrients.

Factors Impacted by a Change in Rainfall IntensityThe WEPP model should be well-suited for

simulating or accounting for an increase in rain-fall intensity. Because the model is process-based, a change of rainfall intensity should beable to be simulated. If measured breakpointprecipitation intensity is available, the effects ofthis storm intensity could be inputted into themodel using proper format and units.The out-put could then be easily read, either for an indi-vidual storm or for the time period for whichrecords are available. If non-breakpoint data areavailable, then simply changing the amount ofrainfall on a given day is all that is necessary. Therainfall amount will be automatically disaggre-gated in a statistically representative mannersimulating a single-peaked storm.

If rainfall intensity is to be modified but notthe rainfall quantity, then changing the ip (peakintensity index) parameter in the input file wouldbe necessary to simulate a storm with the desiredpeak intensity.

Soil and Water Conservation Society 945 Southwest Ankeny Road Ankeny, Iowa 50021-9764

phone 515-289-2331; [email protected]; www.swcs.org

®

conservation implications of climate change: soil …precipitation;two primary conservation...

Documents