A Pilot to Evaluate Water Quality Models For Future Investment

A Pilot to Evaluate Models for Future Investment
Subject Area: Water Quality

Prepared by the Water Resources Working Group (WRWG)
(National Water and Climate Center (NWCC), National Water Management Center (NWMC), Watershed Science Institute (WSSI))
of the Natural Resources Conservation Service (NRCS) Conservation Engineering Division, Washington, DC

In FY00 the staffs of the Water Resources Working Group (WRWG) undertook a pilot to evaluate the wide range of water quality models currently being used or developed in an effort to insure that agency investments are strategic, forward-thinking and appropriate for technical needs in carrying out the NRCS mission. The pilot was done in cooperation with the office of the Deputy Chief for Science and Technology.

It is anticipated that the process used in this project can serve as the basis for other NRCS technical groups as they conduct similar reviews of other computer-based models used in resource management. This paper summarizes key aspects of the water quality pilot.

A team was assembled consisting of five members representing the three entities comprising the WRWG (see Footnote 1). The directors of the three entities served as the sponsors for the team and approved/reviewed various aspects of the effort as it progressed. The team designed a 5-step process for conducting the pilot:

Step 1: Identify Characteristics of Models that Pertain to the Selected Resource Issue

(Compile a list of general features and technical attributes, e.g., model name, brief description, language, scale, inputs and outputs.)

Step 2: Inventory and Record Information

(Identify models used by NRCS that are existing, under development, or envisioned for the future. Populate the table created in Step 1.)

Step 3: Screen the Models for Selected Criteria

(Identify the criteria for the current analysis. Analyze the table of information (e.g., scale, subjective, qualitative, or quantitative to name a few.)

Step 4: Evaluate Models Selected in Step 3

(Consider additional decision factors such as: usability, utility, validation, verification and data availability.)

Step 5: Prioritize Models and Develop the Technology Transfer Support Plan

(Decide which models to operate, develop and expand. Commit the necessary resources. Obtain appropriate Investment Review Board and other coordination.)

In Step 2, the team identified and attributed 30 models; in Step 3, this list was edited to 14 after careful and comprehensive analysis. The team acknowledges that the inventory is limited to those with some connection to or use by NRCS. A web search of water quality models, for example, would turn up more models not on the list, but the team decided that a more limited inventory was appropriate since the primary purpose for the pilot was to target investments. It is important, however, that NRCS technology specialists remain cognizant of other efforts, both domestic and international, so that the agency remains in touch with the state of the art.

• Usability of the current suite of Water Quality (WQ) models is not well suited to a direct field office activity per se. The team found that water quality models require a significant amount of input data preparation and require a certain level of expertise to apply them meaningfully and to interpret the results properly. This time and expertise will not usually be available at the field office level. However, the output of some models (prepared by state office specialists) could be appropriate for inclusion in the Field Office Tech Guide (FOTG) as a tool in the form of a chart, table or graph.

• NRCS cannot expect to have one or two simple water quality models that can be an adequate basis for all land management and conservation practice decisions. Some situations in the field will remain where decisions can be made with simple methods that do not require extensive analysis or sophisticated computer-based tools. Increasingly, however, many WQ issues and decisions must be addressed at the watershed scale rather than the point or field scale since complex processes and spatial interactions are involved.

• NRCS's efficient integration and use of various new technologies/procedures can best be achieved by establishing a modular modeling system that is at a high level of interoperability and application than the array of unique and specialized tools that we now have. A modular modeling system will allow the agency to adapt and use new technology more effectively.

• Models require an investment in back-up support resources. Models are complex, and there are many pitfalls and places for error. The application of a model at any NRCS office requires attention to detail and proper usage if it is to correctly identify or characterize water quality problems and source areas. Coaching, technical support or oversight is areas often neglected as NRCS deploys models for use.

• NRCS needs an identified cadre of peers to work with model developers (especially those at Agricultural Research Service (ARS)) to sort through modeling activities, critique the equations and algorithms used in the models and in general, ensure that the agency obtains models that meet its needs. There are many independent WQ modeling efforts within ARS, other government agencies and universities. Most of these efforts are designed to meet the objectives of research scientists and not necessarily the needs or goals of NRCS. Many of the models have 'components' that are right for NRCS, but there is no one model or tool that would contain all of the 'right' modules for the agency.

• NRCS needs an identified cadre of specialists to provide in-house technical consultation regarding model use and on-site assistance for increasing local capacity for operating water quality models—whether or not the proposed modular system materializes.

• Water quality models in use today by NRCS generally produce results that are most useful for comparative rather than predictive purposes. Models are physically based. They use mathematical equations and algorithms to define complex natural relationships. As the complexity of natural relationships increases, the ability to mathematically "model" them decreases. They can best be used to compare one alternative to another and provide a mechanism to qualitatively rank one alternative against another; however, they should not be used for present or future predictions. Predictive models are increasingly available, but these complex simulations require substantial input data.

The team and the sponsors recommend two requests that constitute a two-pronged approach for future investment in water quality models for FY01. Both proposals are included in budget request for FY01.

One request was submitted in the FY01 budget proposal by the National Water Management Center so that staff can continue to provide direct assistance and support to states using models that may ultimately be phased out in terms of agency investment. Funds are to be used to:

Continue technical support on a few models that are currently in use by NRCS state offices. This item will fund up to four NRCS/ARS/University teams to meet as needed and develop technology transfer plans to insure the use of pollutant loading water quality models are incorporated into the NRCS planning process.Cost: $50,000.

The National Water & Climate Center submitted another request in the FY01 budget proposal so that the Water Science & Technology team and Climate Center modelers could lead the effort. Funds are to be used to:

Hold a symposium to discuss and agree on a process/procedure for building the framework for a modular system of models for future NRCS investment. It is envisioned that the framework would allow developers to prepare one or more elements or modules that, when combined with other modules, could produce a tool to meet a specific need. This approach encourages the "standardization" approach for having an adopted hydrology module, soils module, etc. The work at the symposium, and after, would consist of identifying the appropriate "modules" in existing models and modifying them to fit interchangeably and interoperably into a common water quality modular framework.Cost: $65,000 .

These requests, while submitted under different Water Resource entities, represent the consensus and collective opinion of the NWCC, NWMC, WSSI and the Conservation Engineering Division. It is clear to the leaders and staffs that the NRCS and its partners will continue to suffer from an uncoordinated approach to the development and use of water quality models if changes are not immediately forthcoming. Specifically, the agency will face the following problems:

1. efficient use of modern technology will continue to be cumbersome
2. the watershed approach to water resource problems will continue to be more complex and expensive than it needs to be,
3. existing models will continue to compromise our technical expertise as we will increasingly be hampered by our inability to simulate water quality processes,
4. the agency will be unprepared to integrate new analyses tools without extensive rewriting and expensive modification to existing programs.

The NRCS has long used models developed by our United States Department of Agriculture (USDA) partners and the university system to evaluate the effects of land management on water resources. These "models" or representations of the natural system are often complex, computer-based tools requiring a great deal of expertise to operate them and correctly understand the results of the analyses.

In 1999 the NRCS Water Resource Group - National Water and Climate Center (NWCC), National Water Management Center (NWMC) and the Watershed Sciences Institute (WSSI) - agreed with the Director of NRCS’s Conservation Engineering Division to undertake an evaluation of the current state of the art in water resource models using water quality models to test the process. In a December 1999 meeting of the NRCS Water Resources Group, a five-step model evaluation, development, and support process was prepared. A copy of the process matrix is enclosed as Appendix A.

The NRCS Water Resources Group assigned a team to perform steps one and two of the process. The team members were William Merkel and David Garen with the NWCC, Barry Rosen with the WSSI, and David Moffitt with the NWMC. Donald Woodward with the Conservation Engineering Division and Greg Johnson with the NWCC were asked to serve as advisors to the team, and Ray Riley fully participated with the team in the role as an alternate from the WSSI.

The discussion that follows will provide some background describing NRCS (Soil Conservation Service (SCS)) involvement with water quality models and link to the current effort directed by the Water Resources Group.