Simulation Pollution Loading With Water Quality Models

INTRODUCTION

One of the vexing problems facing conservationists is evaluating the effectiveness of conservation measures applied as part of a resource management plan. What are the environmental changes that will take place with the practices? Do the benefits of the practice in qualitative terms justify the cost?

Not easy questions to answer! From a water quality perspective, the questions can be answered to some extent by monitoring over time. Then the question becomes; “How much can I afford to spend for monitoring?” Even the most basic set of water quality parameter analyses can cost tens to hundreds of dollars per sample, not including the cost of collecting and preserving the sample, recording information about the site when the sample was drawn, and transporting the sample to the place of analysis. A single sample will provide information at the time the sample is collected, but the single sample, like a snapshot, only tells a limited story. Many water quality monitoring projects have wound up costing nearly as much as the cost of practice application, and then yielded only limited information.

Physical process water quality continuous simulation models were developed starting back in the mid-1970’s to provide a methodology to look at the long term impacts of agricultural activities on water quality. The term “continuous simulation” refers to a model that performs repetitive calculations on a pre-set time period, normally daily. The regular time step allows the use of rainfall and other climatic variables commonly found in published records. Models validated at a variety of sites around the world are being used to compare one practice against another and to gain an appreciation of the magnitude of changes in pollutant loadings conservation practices can have. The results of historical simulations can be compared to available information to “calibrate” the model, and then the model used to simulate time periods or conditions not supported by data.

The operative terms are “comparison” and “pollutant loading.” Water quality models don’t actually provide “water quality” information, at least not from a water body or aquifer perspective. What they do is provide an estimate of pollutants delivered to the edge of a field or to some point in a watershed. The assimilation of the pollutant by a water body is another issue and is highly variable. The widely used term, “water quality model,” will be used in the following discussion with the reader understanding they are actually pollutant loading models.

Most of the current models are still not “predictive” in nature, but rather are an indicator of the magnitude of the impact. These comparative analyses allow the modeler to look at one alternative versus another or to look at an alternative compared to the present or baseline conditions. As physical processes such as water movement, soil erosion, and chemical transformations are better understood and can be represented in mathematical terms, the predictive capabilities of water quality models will increase. In addition, water quality models are subject to the “garbage in - garbage out” concept. The accuracy of the model results will directly reflect the quantity and appropriateness of the input data. As more data is used to reflect the field situation, the better the results will be.

FIELD-SCALE MODELS

Field-scale models normally provide estimates of pollutant loading at the edge of field and bottom of root-zone. Some models require the user to define the shape of the field, others don’t. The term “field- scale model” is also applied to models that consider pollutants at a point in the field rather than the field as a whole. In most cases the best the field-scale models do is consider a field-sized watershed.

Most field-scale models consist of components or modules that consider hydrology, erosion-sedimentation, and chemical fate and transport. These are formally linked together in computer code or run as individual modules with the output of one module used as part of the input into the next. Chemical consideration include sediment and water phases of both nutrients and pesticides. Commonly, only nitrogen and phosphorus are included in nutrient considerations. In the list of field-scale models are models such as MWASTE that consider a single parameters such as pathogens.

In the early 1990’s, the Natural Resources Conservation Service (NRCS) (then Soil Conservation Service (SCS)) began a comprehensive review of available modeling techniques. From the many models available, both field and watershed-scale, NRCS selected five models for detailed examination and designation as “approved” for NRCS use. The designation of certain models as “approved” does not mean other models cannot be used, only that when other models are used, the results of the simulations should be viewed with a certain degree of skepticism. The examination included an assessment of technology imbedded in the models as well as states assessing the usability and utility of the models. After the reviews were complete, the NRCS negotiated changes in the models to accommodate the recommendations of the NRCS teams. The three field- scale models chosen for review were the Nitrogen Leaching and Economic Assessment Package (NLEAP) , Ground water Loading from Agricultural Management Systems (GLEAMS) , and Erosion/Productivity Impact Calculator (EPIC). These models were studied in depth, and after changes were made, received the NRCS “approved” designation.

WATERSHED-SCALE MODELS

As with the field-scale models, the watershed-scale models traditionally provide information about the transport of sediments, nutrients, and pesticides. As will be discussed in the section on model use, the watershed-scale models require less site specific information than do field-scale models, and by-and-large provide less site specific information. Predominately the watershed-scale models focus on surface water issues, with little information about root-zone or below the root-zone mechanisms. An exception is Hydrocomp’s Hydrologic Simulation Program - Fortran (HSPF) model developed for Environmental Protection Agency (EPA).

The two watershed-scale models chosen for evaluation by NRCS teams were is Simulator for Water Resources in Rural Basins-Water Quality (SWRRBWQ) and is Agriculture Non-Point Source Pollution Model (AGNPS). Both of these models are presently “approved” for NRCS use. SWRRBWQ is a continuous simulation model using the concept of homogeneous sub-watershed areas as the basis for representing the watershed. AGNPS, on the other hand, is a single event model (the user describes the storm) utilizing cells as the basis for model input. The AGNPS user specifies the cell size in acres.

MODEL USE

The traditional thinking with water quality models is that the use of field-scale models provides an intense look at small areas; whereas watershed-scale models allow a general look at a larger area and blends the impacts of activities on specific fields. The physical processes modeled in field-scale models are generally more detailed than in their watershed-scale counterparts, allowing for the more detailed examination.

Field-scale models require a great deal of information about the surface and sub-surface of the area being modeled. Site specific details of climate, soils, plant growth, chemical application and management are translated by the model into the water quality and quantity values at the edge of the field or the bottom of the root-zone. In field-scale modeling the site being examined is normally represented by one soil, one plant community, and one set of management practices. This does present a problem where the field contains equal areas of multiple soils or multiple crops (such as strip cropping). In these cases the user must decide what is the predominate soil or crop, or in some cases, what combination of soil and crop will best represent the problem situation requiring the modeling effort.

If using a field-scale tool to look at a watershed area, the modeler begins to build a list of typical scenarios within the watershed area. For example, the modeler would simulate various crops on various soils using typical management practices within the watershed. The more variation in the watershed area, the larger the number of scenarios must be simulated for a good representation. The simulated results can be compared to each other (no-till corn vs. corn with conventional tillage for example) for the predominant soils, and a ranking developed for the Field Office Technical Guide (FOTG) relating land use, soils, and management practices.

The field concept is normally masked in watershed-scale modeling in favor of land uses or groups of land uses (pasture vs. cropland for example). The modeler must make some assumptions about what land use, soils, management, climate, etc. best represent the watershed or sub-watershed areas when preparing input data sets. The skill of the modeler in recognizing the best representation of the watershed or sub-watershed during data entry determines the value of the simulation results.

The advent of the interface between Geographical Information Systems (GIS) technology and water quality models has resulted in products such as the Hydrologic Unit Water Quality Tool (HUWQ) , and provides the opportunity to blend field-level data with watershed-scale modeling technology. Individual field and soil boundaries are identified by GIS layers, and descriptive attributes assigned to each of the field and soil units. The interface then calculates the most appropriate model input data for each cell or sub- watershed based on the attributes assigned for each field and soil within the cell or sub- watershed. This eliminates the necessity for assuming or manually determining the most representative input values, which often can introduce erroneous data into the modeling process. The HUWQ Tool and the GIS interface for Soil and Water Assessment Tool (SWAT) also provide graphical representations of the model outputs not provided by the original models.

< Back to Water Quality