Integrated modeling of nutrient and sediment sources and sinks across spatial scales

Large watershed-scale models:

We will use 2 models at the scale of the entire upper Susquehanna River basin: the SCOPE/NANI model and the Regional Nutrient Management model (ReNuMa). We will explore how the insights and output from several smaller scale models can interact with these large watershed-scale models.

The SCOPE/NANI model is a simple mass-balance model for N that compares landscape sources of N to riverine N fluxes. Originally developed for large regions, such as the combined watersheds of the North Sea or the northeastern U.S., or the entire Mississippi River basin (Howarth et al. 1996), it has subsequently been applied to watersheds of the scale of the Susquehanna both in the U.S (Boyer et al. 2002) and in Europe (Humborg et al., unpublished). Despite its simplicity, a comparative analysis of many models demonstrated that the SCOPE/NANI model is among the best in terms of error of prediction and assessment of N source determination (Alexander et al. 2002). The model has been employed primarily to evaluate average N fluxes over periods of 6 to 10 years, but McIsaac et al. (2001) used a modification of the approach to accurately predict year-to-year variations in N flows in the Mississippi River that are associated with climatic variation. We will test this approach for the Susquehanna, in an effort to better determine the influence of climate on overall N fluxes, and the relative importance of climate effect on different N source terms (Howarth et al., in press.). Climate variability and climate change are likely to have a profound effect on nutrient transport from landscape to rivers to coastal ecosystems, but there is great uncertainty as to the detailed responses expected (Boesch et al. 2000; Scavia et al. 2002).

ReNuMa is a large-watershed-scale model designed to allow planners and other stakeholders to explore scenarios for reducing N fluxes from the landscape. More information about this model

Integration of different models, and across different spatial and temporal scales:

Limited funding exists from the USDA Hatch program to begin to apply ReNuMa to the Susquehanna basin. This additional funding will enhance our ability to parameterize the model to the Susquehanna. We also propose to use output from the SCOPE/NANI model to improve the ability of ReNuMa to predict the consequences of climate variability and climate change on delivery of N. Also, we will use insights gained from plot- to farm-scale models to better parameterize the responses of ReNuMa to agricultural practices (see discussion below on these models). Both the SCOPE/NANI approach and ReNuMa will be ground-truthed against the accumulating body of data on fluxes of water, N, P, and sediments from the upper Susquehanna River basin being collected by the USGS, the New York State DEC, and others.

Many models exist which address fluxes of nutrients at the scale of plots to farm fields to whole farms, including several models developed and used by Cornell faculty and staff. These often have more complexity and spatial reality than do large watershed models such as SCOPE/NANI and ReNuMa. On the other hand, their complexity and detail tend to make them difficult to scale up to give reliable estimates of watershed-scale nutrient fluxes (NRC 2000). Using one scale of modeling to inform modeling at a different spatial scale, although challenging, is highly desirable.

We propose to facilitate such an interaction of modeling efforts across scales in the Susquehanna River basin. One goal is to improve the utility of ReNuMa, as stated above. Beyond this, we believe the cross-fertilization of ideas may lead to improved models at smaller scales as well, and to an improved understanding of the processes that are critical in determining the fluxes and sinks of N and P (and where in the landscape such processes may be most important). One approach to crossing this scale divide includes statistical comparisons of model outputs when models are subject to similar input perturbations, such as extreme weather events. Another approach is to use the model research at smaller scales to better identify and describe the critical processes that may control the overall behavior of large watersheds. For example, preferential flowpaths may be characterized by their relationship to soil types. Various categories of land and crop management may be summarized in terms of their effects on hydrology, soil loss, or chemical processes. This level of information can be considered a "typology" or classification system derived from detailed process studies, and can in turn be used as a basis for simplified models or "decision support tools" at larger scales. An approach like this has been taken in Sweden to develop the SOILNDB system, based on the models SOIL and SOILN.

Collaborative modeling funding opportunity:

Given the large number of smaller scale models of potential value to the overall goals of the project, we propose to have a competitive process for selecting both for the model studies that may offer the best opportunity to scale up insights, and for researchers who are most open to undertaking this challenging endeavor.