Modeling land use patterns and water quality: An evaluation of the pySPARROW model

Abstract

Modeling the effects of land use and land cover changes on water quality is important for watershed managers to better understand how human modifications to land surfaces may alter stream nutrient loads. One model available to resource managers for this purpose is the U.S. Geological Survey's SPARROW (Spatially Referenced Regressions on Watershed Attributes) model. SPARROW estimates total nitrogen and total phosphorus loads for watersheds by relating water quality information to nutrient sources, land-surface characteristics, stream connectivity, and downstream travel time. This project evaluates the pySPARROW model, which is an application of SPARROW written in the Python programming language for North Carolina's non-tidal stream network. By analyzing estimated nutrient loads of the Falls Lake subbasin under current land uses, this project assesses how well pySPARROW predicts the long term mean total nitrogen concentration. A regression analysis of the observed versus predicted total nitrogen concentrations shows that pySPARROW most likely needs to be recalibrated to improve its accuracy. The model is also used to assess watershed impacts of a development scenario under which forests and agricultural lands are converted to urban uses. Under this scenario, the total nitrogen loading of the Falls Lake subbasin increases and the loading of most catchments which experienced some development also increases. With the population of the Falls Lake subbasin expected to increase by 50 percent from 2000 to 2025, it is especially important that watershed managers have tools, such as pySPARROW, that may be used to predict the impact of land use changes on water quality in this region.