The Bay Program is undertaking a wholesale update of its Watershed Model. Known as Phase 5, the revised model is expected to provide the most accurate predictions yet of the amount of nutrients entering the Bay, and where they are coming from.

The Watershed Model has been the yardstick by which the Bay Program has measured nutrient reductions since the 1980s. Although it has undergone a number of refinements over time, its results in the past have often differed from observed water quality monitoring.

Part of the difference is not caused by the model itself, but by the information going into the model. In some cases, the effectiveness of nutrient reduction actions have been overestimated, which in turn causes the model to overestimate nutrient reductions. Also, the number of nutrient control practices may be overestimated, or underestimated, from place to place, affecting model estimates.

But the model itself is also limited by the amount of information it can handle. The current model, for instance, divides the watershed into large segments—each of which may contain a variety of different soil types and land uses—and has to average all of that information together, which can decrease accuracy.

The new model will have improved spatial accuracy in nutrient and sediment estimates by having more segments. While the current model divides the 64,000-square-mile watershed into 94 segments, the new model will have nearly 10 times as many—931 segments covering land and rivers throughout the basin. In most cases, each county will be represented by its own segment in the model.

The new model will include more up-to-date information on land uses, derived from new satellite imagery, so it can estimate nutrient exports from more types of land than before.

The model will incorporate more spatially detailed precipitation information, which is important for estimating the amount of nutrients washed off the land—heavy downpours flush more nutrients into the water than slow, soaking rains.

Reservoirs can have major impacts on the transport of nutrients, generally resulting in lesser amounts of nitrogen and phosphorus downstream. The current model only simulates 6 reservoirs; the updated model will simulate 40.

The existing model used only 16 monitoring stations as the basis for its river flow simulations. The new model uses data from 290 flow stations. And it will have more than 70 water quality stations where model simulations and actual water conditions can be compared—up from a couple dozen sites today. That means there will be more opportunities to compare model results with real-world conditions, which will help improve the accuracy of modeled estimates.

Some changes are still being explored. In the real world, much of the nitrogen moves through the groundwater before it reaches streams, a journey that can take years, thereby delaying the benefit of some nutrient control actions. The current model does not account for that delay and therefore counts those nutrient reductions before they show up in the Bay. Officials hope to account for that “lag time” in the new model if a technique for doing so can be developed.

One issue will likely arise from this. As model estimates become more accurate, the estimated nutrient loads from various basins—and land uses—will likely change. The new model will be used when the Bay Program re-evaluates its nutrient reduction goals in 2007, at which time it’s likely that tributary strategies will need to be updated.