CONSERVATION PLANNING FOR CONSTRUCTION SITES D. C. Yoder, J. S. Tyner, J. D. Balousek, J. C. Panuska, J. R. Buchanan, K. J. Kirsch, J. P. Lyon

ABSTRACT. Although conservation planning for agricultural lands has evolved to its current relatively stable form over many decades, conservation planning for construction sites is still in its infancy. This project drew on the resources of various agencies and researchers to develop a conservation planning tool specifically geared towards meeting construction site planning needs in Wisconsin, basing it on the Revised Universal Soil Loss Equation, version 2 (RUSLE2). The project began by deconstructing the planning process itself to determine possible approaches and critical elements, and used those pieces to build a rational new approach to construction site conservation planning. Although no changes were required to the erosion or sediment delivery calculations in RUSLE2, this new tool required substantial changes to the database, the interface, and how the results were presented and packaged. The lessons learned in this effort should be instructive to both a general discussion of the construction site planning process and to attempts to develop other tools that meet this need. Keywords. Conservation, Construction, Erosion, Sediment.

E

fforts to manage erosion from agriculture have been ongoing for centuries (Lal, 1998), and modern scientific tools to predict agricultural erosion have been available and widely applied within the U.S. for decades (Wischmeier, 1976), so the science and tools to estimate and manage agricultural erosion are relatively ma‐ ture. Recently, however, more interest has been given to man‐ aging erosion on and sediment yield from urban construction sites. Instead of falling under the jurisdiction of a large feder‐ al umbrella such as that held by the USDA, urban erosion is managed by multiple levels of federal, state, county, and city regulations attempting to maintain water quality as mandated by the Clean Water Act (USEPA, 2006). These ordinances vary greatly, are typically qualitative in nature, and do not necessarily ensure that the sediment delivery is below some predetermined threshold deemed suitable to all stakeholders. This article describes an attempt to apply the conservation

Submitted for review in February 2007 as manuscript number SW 6908; approved for publication by the Soil & Water Division of ASABE in July 2007 as a contribution to the ASABE 100th Anniversary Soil and Water Invited Review Series. The authors are Daniel C. Yoder, ASABE Member Engineer, Professor, and John S. Tyner, ASABE Member Engineer, Assistant Professor, Department of Biosystems Engineering and Soil Science, University of Tennessee, Knoxville, Tennessee; Jeremy D. Balousek, ASABE Member Engineer, Urban Conservation Engineer, Dane County Land Conservation Division, Dane Co., Wisconsin; John R. Panuska, ASABE Member Engineer, Natural Resource and Bio‐Environmental Engineer, Department of Biological Systems Engineering, University of Wisconsin, Madison, Wisconsin; John R. Buchanan, ASABE Member Engineer, Associate Professor, Department of Biosystems Engineering and Soil Science, University of Tennessee, Knoxville, Tennessee; Kevin J. Kirsch, Water Resources Engineer, Wisconsin Department of Natural Resources, Madison, Wisconsin; and James P. Lyon, IT Analyst, Department of Biosystems Engineering and Soil Science, University of Tennessee, Knoxville, Tennessee. Corresponding author: Daniel C. Yoder, Department of Biosystems Engineering and Soil Science, University of Tennessee, 2506 E. J. Chapman Dr., Knoxville, TN 37996‐4531; phone: 865‐974‐7116; fax: 865‐974‐4514; e‐mail: [email protected].

planning tool RUSLE2 to such construction settings, which required changes not only in the technology, but also in the very approach to conservation planning.

BACKGROUND Conservation planning for agriculture has become so en‐ grained that it is generally taken for granted. Ideally, planning is done by a well‐trained expert applying standard techniques and technology in helping the producer examine manage‐ ment alternatives. If soil erosion is a primary concern, the technology of choice has usually been the Universal Soil Loss Equation (USLE) (Wischmeier and Smith, 1978) or more recently the Revised Universal Soil Loss Equation (RUSLE) (Renard et al., 1997). The technical expert visits the field to collect topographic and other necessary informa‐ tion and then uses USLE/RUSLE to estimate the average annual erosion rate, which is compared to the soil loss toler‐ ance (T value) (discussed in Renard et al., 1997, pp. 12‐13) to determine the adequacy of the proposed management prac‐ tice. If the USLE/RUSLE result compares poorly with the target T value, then the expert and manager try another man‐ agement alternative. The details of this conservation plan‐ ning approach have often been specified by regulations in the USDA farm program for which the producer was applying. Such specifications usually required that the plan meet a tar‐ get soil loss or soil loss reduction goal, and included occa‐ sional spot‐checking field verification of compliance with the plan. This planning process has evolved with new technologies (such as RUSLE2; Foster, 2004, 2005), new conservation targets (such as the Conservation Security Pro‐ gram; NRCS, 2007a), and new technical expert guidance (such as Technical Service Providers; NRCS, 2007b), but the general approach has remained the same, with a technical ex‐ pert helping the producer examine alternatives by comparing the results of technology‐based estimates. How does conservation planning for construction sites compare to this process? In at least one general way, it is simi‐

Transactions of the ASABE Vol. 50(5): 1613-1618

E 2007 American Society of Agricultural and Biological Engineers ISSN 0001-2351

1613

lar. As in agricultural settings, there are really two customers for the conservation planning tool. The first customer is the producer or planner, who in each case wants to examine the broadest possible range of alternatives so as to pick the one that minimizes cost. The other customer is society at large, which has a vested interest in maintaining the quality of local and regional soil and water resources, and in minimizing po‐ tential damage to public stormwater systems. The planning process seeks a compromise that satisfies both parties. Conservation planning for construction sites, however, is different from that for agricultural lands in that it is typically not well defined, or is defined by a set of codes with little quantitative basis. Planners producing Storm Water Pollution Prevention Plans (SWPPPs) are attempting to meet require‐ ments that vary between states, and often between counties or municipalities within a state. These requirements can range from very specific codified target values (e.g., specific sediment load or turbidity values) to a menu of best manage‐ ment practices that the planners can apply without significant justification. These plans are reviewed to varying degrees of detail by regulatory agency personnel, and once the plan is implemented regulatory field inspectors have differing levels of authority to demand changes. In classes and training ses‐ sions led by several of the authors, all of the parties involved (planners, reviewers, inspectors, and developers) com‐ plained about the lack of consistency and specificity. Plan‐ ners are never quite sure when a plan is satisfactory, either because there is no specified target or sometimes because the target is so unrealistic that it is routinely ignored. Reviewers and inspectors are forced to make subjective judgments of adequacy, leaving themselves open to criticism and second‐ guessing by the planners and developers, and sometimes leading to appeals of the review. These personnel all ex‐ pressed a desire for a single tool that could definitively deter‐ mine for all parties whether a plan is adequate, as USLE/RUSLE does for agriculture. In addition to the lack of plan definition, construction site management schemes are also often very different from those for agriculture. Construction sites are routinely left com‐ pletely bare for extended periods, the massive land‐shaping activities usually leave none of the original protective bio‐ mass on or near the surface, and even the soils differ in that they are typically exposed subsoils or topsoil‐subsoil mix‐ tures. These unique characteristics must be taken into ac‐ count in any effort to develop a construction site conservation planning tool. A project sponsored by the USEPA, directed by the Wis‐ consin Department of Natural Resources (WIDNR), and in‐ volving the Dane County Land and Water Resources Department (Dane Co.) and the Department of Biosystems Engineering and Soil Science at the University of Tennessee set out to address this need for a construction site conserva‐ tion planning tool. The effort followed previous use by Dane Co. of a USLE‐based tool (Balousek et al., 1999) but made use of the advanced capabilities of RUSLE2, with additional modifications that would ease its use for construction sites. RUSLE2 was chosen by the WIDNR and Dane Co. for the following reasons: (1) RUSLE2 is already actively being used by the USDA‐NRCS for agricultural planning, with an estimated 3000 to 5000 uses per day, so it is a relatively robust and mature technology; (2) many of the database elements (climates, soils, and some vegetation and management de‐ scriptions) required by RUSLE2 are already available from

1614

the USDA‐NRCS (NRCS, 2007c); (3) RUSLE2 can easily model most of the additional construction site activities not already in the NRCS database, including management prac‐ tices such as mulches, blankets and vegetations, devices or structures such as permeable barriers (e.g., silt fences, straw bales) and sediment basins, and combination techniques such as vegetative filter strips; (4) RUSLE2 enables simple and quick comparisons between multiple management alterna‐ tives, allowing planners to perform rapid cost‐benefit analy‐ ses; (5) the flexible RUSLE2 user interface can be modified to more appropriately represent construction sites and how their planners approach things (icons, terminology, look and feel, etc.); and (6) RUSLE2 contains built‐in protections on what various users can see and change, including material found in the database, thereby enhancing data security. No changes in the RUSLE2 science described in Foster (2004, 2005) were required to develop this new tool; the changes were rather in the packaging and the presentation of the re‐ sults. This article has the following objectives: (1) to describe the general approach selected for the erosion planning pro‐ cess; (2) to explain how that approach was implemented for this project, including RUSLE2 modifications; and (3) to dis‐ cuss needs for further work in construction site conservation planning that became apparent during the effort. The analysis of the planning process is specifically geared towards the use of RUSLE2, but similar issues would arise for any compara‐ ble erosion model. The approach, the implementation effort, and other lessons learned during this process should be instructive and useful for any conservation planning process, particularly for those related to construction sites.

THE PLANNING PROCESS To properly investigate construction conservation plan‐ ning, it is helpful to examine several currently successful planning schemes. One example is the USLE for agricultural planning, as described previously. The curve number meth‐ od, also known as the SCS direct runoff method (SCS, 1993), is another planning tool, used to estimate total runoff volume from small watersheds. This technique is described in detail in most conservation texts (e.g., Haan et al., 1994; Schwab et al., 1993), is relatively simple, and is very familiar to most resource professionals, so it will serve as an instructive exam‐ ple of the planning process. The SCS curve number method estimates total runoff vol‐ ume and is commonly used to size detention basins. This method requires only two inputs: a total storm rainfall depth, and the curve number, which describes the propensity of the watershed to produce runoff. A classical risk‐analysis ap‐ proach based on the curve number method would look at the potential distribution of all the input values (in this case, the storm depth and curve number), would use values selected from those distributions to calculate the distribution of out‐ puts (in this case, the runoff volumes), and would estimate for each input/output set the total cost, including both the cost of the design and associated damage to public infrastructure, health, and safety (Haar, 1987). Once the costs were all in the same terms (usually dollars or another currency), an opti‐ mization function could be used to minimize total costs for each design, allowing selection of the most cost‐effective. Although attractive in principle, such an approach is very dif‐

TRANSACTIONS OF THE ASABE

ficult in practice. First, associating costs with human health and safety and the many other factors involved is very impre‐ cise and controversial. In addition, uncertainty in the inputs' relationships and in the scientific model creates increasing potential error as the uncertainty cascades through the sys‐ tem, leading to an optimized output in which the designer may have little confidence. An alternative design approach uses the same initial steps to generate an output distribution (volume of runoff) but rath‐ er than converting this into costs, simply compares the output distribution to some predefined performance criterion, or measure of success (Haar, 1987). This allows for estimation of the risk of the criterion being exceeded, and thus of the de‐ sign being defined as failing. For example, in the case of a detention basin designed using the curve number method, the performance criterion would be the volume of water that the proposed design could hold, so this approach provides the probability of exceeding that volume. A third design approach represents the most common usage of the curve number. This begins with single values picked from the input distributions to give a single‐valued output, which is then compared to the performance criterion. In this approach, the user chooses a best estimate of the curve number for the area and then selects a design rainfall depth from the input rainfall distribution using a local intensity‐ duration‐frequency (IDF) curve from TP‐40 (Weather Bu‐ reau, 1961) or some similar source. This rainfall selection requires choosing the design rainfall duration and recurrence interval (also known as the return period), which defines the location of the selected value in the rainfall depth distribution curve. By convention, a 24‐hour design rainfall duration is often selected for basin design, and the recurrence interval is based on the presumed cost associated with failure, defined as the detention basin not being able to contain the design storm volume. While a 5‐year 24‐hour design storm might be appropriate for a small area with a non‐sensitive downstream area, a 25‐year 24‐hour design storm would be a more ap‐ propriate design storm for a larger area with a sensitive down‐ stream area. The 25‐year 24‐hour storm event is larger, so it requires a more conservatively sized and therefore more ex‐ pensive basin to limit the risk of failure, since failure would represent a larger cost to society. This technique of setting de‐ sign recurrence intervals is the most common method of im‐ plementing stormwater design requirements. It qualitatively reduces the risk of failure for larger projects and minimizes the cost to developers for smaller projects. One weakness of this approach is the implicit assumption that the return period of the runoff event is the same as that of the storm, which is often not the case (Haan et al., 1994). A fourth design approach often used in engineering prob‐ lems takes as inputs the average values rather than using the input distributions, and takes the resulting single output value as the design parameter, but multiplies that by a factor of safe‐ ty (Hyman, 2002). The size of the factor of safety is based on both the cost of failure and a qualitative analysis of the uncer‐ tainty in the design parameters. For example, the factor of safety for a wood design is often greater than that for steel, because wood strength properties tend to be more variable. From a conceptual basis, this approach takes some of the pro‐ babilistic elements of the previous approaches and incorpo‐ rates them in the design through the performance criterion rather than through the inputs.

Vol. 50(5): 1613-1618

All four of these design approaches are valid planning ap‐ proaches in common use, and all make use of three critical planning elements, which include: (1) the performance crite‐ rion, (2) the duration of interest, and (3) the recurrence inter‐ val or return period. The performance criterion is the predefined measure on which the success of the design will be judged. While for a total cost analysis this will be the low‐ est total dollar value, in most designs the performance criteri‐ on will be some physical limit, such as the maximum basin storage volume. The duration of interest defines over what time interval that performance criterion will be judged. For example, in the case of the detention basin, we often use a 24‐hour storm to ensure that the basin will handle a large vol‐ ume of water; a 30‐minute storm on the area might have a higher intensity and thus produce a higher peak runoff rate, but will not yield as great a volume. Finally, the recurrence interval or return period defines the likelihood of the perfor‐ mance criterion being exceeded, although as shown above this can alternatively be subsumed into the performance cri‐ terion through a factor of safety. These three critical planning elements are also present in the example of applying the USLE method to designing a conservation system for an agricultural setting. The perfor‐ mance criterion is the acceptable soil loss tolerance (T value) to which the resulting average annual erosion value (Mg ha-1 year -1 or tons acre-1 year-1) is compared. By definition, the duration of interest is one year, since both the performance criterion and the result have that duration. Finally, use of the average annual value defines the recurrence interval as two years, since this value has a 50% probability of being exceed‐ ed in any year. Note that both the USLE results and the perfor‐ mance criterion are defined based on this same recurrence interval. The following general observations can be made with re‐ gard to the critical planning elements described above: S The elements are interdependent and cannot be sepa‐ rated, so a performance criterion makes no sense with‐ out a specified duration and recurrence interval. S Some performance criteria can be considered as “hard” numbers, such as the basin volume that should not be exceeded. Many other criteria are based on more quali‐ tative judgments or less‐certain models, including such examples as the T values mentioned previously, TMDL model results, or limits set on the basis of biotic integri‐ ty indices. S The performance criterion and associated duration and recurrence interval must fit with the goals of the plan‐ ning process. For example, if the erosion and sediment control measures are designed to stop sheet and rill ero‐ sion, it makes no sense to model a 500‐year storm, as this would likely cause slope mass failure. DEFINING THE PLANNING PROCESS FOR CONSTRUCTION SITE CONSERVATION The planning approaches and critical elements described above should apply to any planning process, but we are most interested here in their application to conservation planning on construction sites. The remainder of this article will con‐ centrate on application of these approaches and elements to the Wisconsin project. Many specific design decisions were made for this test case, often relying on the developers' expe‐ rience and judgment, so we readily admit that other ap‐ proaches and/or parameter values could have been chosen.

1615

We are not proposing that this specific model be followed precisely, but rather that the proper approaches and elements be invoked in some shape or form. The first element defined by the project was the duration of the period of interest, termed the “accounting period,” and defining the period for which the construction planner is re‐ sponsible for controlling sediment delivery from a site. The choice of accounting period is a policy decision that could vary by location and specific water‐quality interests. For this project, the accounting period began with the first soil‐ disturbing field operation and ended with the application of permanent erosion protection, defined as either application of a semi‐permanent non‐erodible surface (pavement, land‐ scape fabric and cover, sod, etc.) or 60 days of growth of a pe‐ rennial vegetation. These 60 days of growth were not allowed to include days when the average air temperature was below 1.7°C (35°F), thereby giving no growth credit for dormant periods of fall‐seeded vegetation. This approach gives the planner an incentive to keep the accounting period short, to reduce erosion and delivery during that period, to plan construction during non‐erosive periods, and to plant cover when it will grow, all of which are good conservation plan‐ ning practices. RUSLE2 was then used to calculate the sediment delivery (Mg ha-1 or tons acre-1) accumulated during the entire ac‐ counting period, representing the best estimate of the ac‐ counting period delivery experienced if the practices were applied on the site next year, assuming average conditions for the climate and associated parameters. This would corre‐ spond to a recurrence interval of two years. Support for this approach came from field practitioners involved in the effort, who expressed more interest in system performance under two‐ to five‐year than under 10‐ to 25‐year recurrence inter‐ vals, as they believed that the extreme nature of the latter might cause failure of many reasonable systems. The estimated sediment delivery during the accounting period was then compared to a performance criterion to mea‐ sure the conservation plan's success. Since it was difficult to scientifically establish a numerical value for maximum al‐ lowable sediment delivery, and since such a standard would contain substantial uncertainty and be difficult to defend, a different approach was chosen. A group of local erosion‐ control field experts used the USLE and simple sedimenta‐ tion approaches to model on typical Dane County construction sites a variety of management practices, includ‐ ing various combinations of mulches and vegetative growth as well as sediment traps and basins (WIDNR, 1994). Any system that reduced sediment delivery by at least 75% was considered a “good” management system. As part of the cur‐ rent effort, these actual scenarios in Dane County were then simulated with RUSLE2. These efforts yielded very consis‐ tent results for the “good” management practices that were modeled, all resulting in about 11.2 Mg ha-1 (5 tons acre-1) of sediment delivery during the accounting period, so this value was established as the performance criterion. This is clearly a subjective criterion, but as described previously, many such criteria are uncertain and subjective. In addition, this approach has the following advantages: S This technique can be characterized as “calibrating” the performance criterion to the model used to estimate plan performance. Models such as RUSLE2 typically have much smaller relative errors than absolute errors. Therefore, with this approach, the errors in setting the

1616

performance criterion and in estimating the plan sedi‐ ment delivery may be seen as largely canceling, pro‐ viding a better result. S This technique ensures that a plan applying “good” ero‐ sion and sediment control systems can indeed meet the performance criterion. One potential problem with us‐ ing a performance criterion set by another method is that it might be discovered later that no real systems can be modeled to meet the criterion. For example, if a wa‐ ter quality model used to set the performance criterion indicated that the maximum allowable period sediment delivery was 1 Mg ha-1 (0.5 tons acre-1), then even the best erosion control systems normally used on construction sites might not be able to meet the target. S A second layer of protection can be added to the sedi‐ ment delivery performance criterion by using a factor of safety. For example, if the runoff enters a sensitive waterbody, regulators could reduce the acceptable cri‐ terion by 1/2 or some appropriate value based on water quality requirements. IMPLEMENTING THE PLANNING PROCESS IN RUSLE2 To implement this planning process, RUSLE2 required several technical modifications, which presumably would also be required of other tools that might be used for this pur‐ pose. The first change was to implement the accounting peri‐ od. Rather than summing daily outputs to get average annual values (as for agricultural planning), the newly defined ac‐ counting period required summing daily outputs only within the specified time window. In making this change, the defini‐ tion of the accounting period was left very flexible, so that other users could define the accounting period differently. From this conceptual viewpoint, the annual value is simply another option for the accounting period definition. Changes to RUSLE2 also allow for requiring that multiple performance criteria be met in order for a plan to pass. For example, RUSLE2 can calculate the maximum daily sedi‐ ment delivery value during the accounting period, which could be compared to some additional criterion established using a TMDL plan. A plan could be required to meet both the accounting period criterion and the maximum daily value criterion. In RUSLE2 agricultural uses, once a management de‐ scription (list of dates and associated operations) is defined, it is usually saved for later reuse, since the way one producer in a region manages corn with spring chisel tillage is likely to be very similar to other local producers. This is much less likely in construction settings, where the timing and se‐ quence of practices is controlled by many factors other than the season. The RUSLE2 interface was therefore changed to make entry of the management description far more evident and direct. This was especially the case for placement and re‐ moval of permeable barriers, which require substantial flexi‐ bility in dates. The RUSLE2 interface was also changed in “look and feel” to better reflect how a construction site plan‐ ner sees things. This included changes in the terminology (e.g., use of “cover materials” rather than “residues”) and in the program icons, text, and other visual representations. The RUSLE2 reporting process was also enhanced for use in the more regulatory environment associated with construc‐ tion site conservation planning. The submission process was made easier by providing a one‐click document generator that creates a non‐modifiable Microsoft WordR file. This file

TRANSACTIONS OF THE ASABE

includes narrative sections entered by the planner, as well as a list of the RUSLE2 inputs and outputs, specifying which in‐ puts were taken from the locked database and which were en‐ tered by the planner. In addition, attached to this document is a full set of the input data for the RUSLE2 calculation, al‐ lowing the plan reviewer to load it directly into RUSLE2 for examination if desired. Finally, use of RUSLE2 required additional database de‐ velopment to model practices not currently available in the extensive NRCS database (NRCS, 2007c). This included es‐ pecially erosion control blankets and mats, but also included other cover applications and such additional practices as track‐walking, which is driving a tracked vehicle up‐and‐ down slope, with the track pads creating a series of small de‐ pressions and ridges roughly on the contour. To understand the effectiveness of these practices, the project included an extensive literature review of erosion on construction site set‐ tings, and the resulting values were used to add descriptions to the practices stored within the database. These practices can be selected by model users for application in their plans, but the descriptions themselves are locked and cannot be modified. All of these changes were deemed essential to the effec‐ tive use of RUSLE2 for construction site conservation plan‐ ning. These changes began with a clear choice of planning approach and definition of the critical planning elements (cri‐ terion, duration, and recurrence) and then included the changes required to effectively implement those elements. These changes were presented in December 2006 to a select group of RUSLE2 novices representing both end‐user and regulatory groups, all of whom received both the approach and the new technology with enthusiasm.

ADDITIONAL NEEDS Of 300 references examined in the literature review for this project, only 60 were deemed to contain sufficient infor‐ mation to allow them to be useful for a USLE/RUSLE‐based sheet and rill erosion approach. Even among those 60, how‐ ever, there was tremendous variability in the sizes of plots, the applied or natural rainfall intensities, the timing of the tests with respect to application, and many other factors. This pointed to a pressing need to establish a standard protocol for the testing of these practices and products, which is currently underway to some extent under the auspices of ASTM. Such protocols evolved over the decades for agricultural research (Lal, 1994), but the construction erosion control community has yet to arrive at a consensus, slowing both scientific and planning progress. There is also a need for construction site planning tools to make easier use of the 3‐D topographic information available in GIS‐based data and planning tools, which would greatly ease conservation planning by allowing RUSLE2 to link di‐ rectly with those tools. Again, there are efforts underway (by USDA‐ARS) to make a future version of RUSLE that can more easily interact with GIS‐based systems, and to develop the GIS‐based tools that will be needed to best make use of that new version.

Vol. 50(5): 1613-1618

CONCLUSIONS Conservation planning for construction sites has often frustrated the planners, plan reviewers, field inspectors, and developers because the planning process itself is not clearly defined nor always evenly enforced. The conservation plan‐ ning process for construction sites is similar to that for agri‐ cultural settings in that it must have clearly defined values for the critical elements, including a performance criterion on which to judge success, and an associated recurrence interval and duration of the period of interest. For the project on which this article is based, the performance criterion was defined by using the planning tool (RUSLE2) to determine the expected sediment delivery for a “good” erosion and sediment control system. The accounting period duration was defined from the first soil disturbance to placement of semi‐permanent erosion‐resistant materials (e.g., pavement or established pe‐ rennial vegetation), and the recurrence interval was set at two years by using average values. Changes to the user interface were made to allow use pat‐ terns more compatible with how RUSLE2 will likely be used for construction site planning. These changes included visual changes and additional tools to enhance use in a regulatory setting, including special mechanisms to lock protected data‐ base elements and to provide reports whose authenticity can be validated. This project also uncovered additional future changes needed to enhance the planning process for construction sites. These included changes to RUSLE2 to make better use of GIS‐based information, but the most significant need is for a standard protocol for construction site based erosion control testing. ACKNOWLEDGEMENTS The material describing the planning process is based on a presentation given at the “Planning for Extremes” work‐ shop (Milwaukee, Wisc., 1‐3 Nov. 2006) sponsored by the Soil and Water Conservation Society, Ankeny, Iowa.

REFERENCES Balousek, J. D., A. Rao‐Espinosa, and G. D. Bubenzer. 1999. Predicting erosion rates on construction sites using the Universal Soil Loss Equation in Dane County, Wisconsin. ASAE Paper No. 992158. St. Joseph, Mich.: ASAE. Foster, G. R. 2004. Draft User's Reference Guide. Undergoing review: currently available in draft form at: www.ars.usda.gov/Research/docs.htm?docid=6028. Accessed 10 February 2007. Foster, G. R. 2005. Draft Science Documentation for RUSLE2. Undergoing review: currently available in draft form at: www.ars.usda.gov/Research/docs.htm?docid=6028. Accessed 10 February 2007. Haan, C. T., B. J. Barfield, and J. C. Hayes. 1994. Design Hydrology and Sedimentology for Small Catchments. San Diego, Cal.: Academic Press. Haar, M. E. 1987. Reliability‐Based Design in Civil Engineering. New York, N.Y.: McGraw‐Hill. Hyman, B. I. 2002. Fundamentals of Engineering Design. 2nd ed. Upper Saddle River, N.J.: Pearson Education. Lal, R., ed. 1994. Soil Erosion Research Methods. 2nd ed. Boca Raton, Fla.: CRC Press.

1617

Lal, R. 1998. Soil erosion impact on agronomic productivity and environment quality. Crit. Reviews Plant Sci. 17(4): 319‐464. NRCS. 2007a. Conservation Security Program. Washington, D.C.: USDA‐NRCS. Available at: www.nrcs.usda.gov/Programs/csp/. Last updated 9 January 2007. Accessed 10 February 2007. NRCS. 2007b. NRCS Technical Service Provider. Washington, D.C.: USDA‐NRCS. Available at: techreg.usda.gov. Accessed 10 February 2007. NRCS. 2007c. RUSLE2 Official NRCS Database. Washington, D.C.: USDA‐NRCS. Available at: fargo.nserl.purdue.edu/rusle2_dataweb/RUSLE2_Index.htm. Accessed 10 February 2007. Renard, K. G., G. R. Foster, D. K. McCool, and D. C. Yoder, coordinators. 1997. Predicting Soil Erosion by Water: A Guide to Conservation Planning with the Revised Universal Soil Loss Equation (RUSLE). Agriculture Handbook 703. Washington, D.C.: USDA. Schwab, G. O., D. D. Fangmeier, W. J. Elliott, and R. K. Frevert. 1993. Soil and Water Conservation Engineering. 4th ed. New York, N.Y.: John Wiley and Sons.

1618

SCS. 1993. Section 4: Hydrology. In SCS National Engineering Handbook. Washington D.C.: USDA Soil Conservation Service. USEPA. 2006. Laws and Regulations: Clean Water Act. Washington, D.C.: U.S. Environmental Protection Agency. Available at: www.epa.gov/r5water/cwa.htm. Last updated 14 July 2006. Accessed 26 February 2007. Weather Bureau. 1961. Rainfall Frequency Atlas of the United States. Technical Paper No. 40. Washington, D.C.: Department of Commerce, Weather Bureau. WIDNR. 1994. The Wisconsin Water Quality Assessment Report to Congress. PUBL‐WR 254‐94‐REV. Madison, Wisc.: Wisconsin Department of Natural Resources. Wischmeier, W. 1976. Use and misuse of the universal soil loss equation. J. Soil Water Cons. 31(1): 5‐9. Wischmeier, W. H., and D. D. Smith. 1978. Predicting Rainfall Erosion Losses - A Guide to Conservation Planning. Agriculture Handbook 537. Washington, D.C.: USDA.

TRANSACTIONS OF THE ASABE