Over the past century, agriculture has kept up with demand from soaring populations by using more intensive growing practices and by converting large swaths of land to pasture. In particular, widespread applications of nitrogen and phosphorus fertilizers have fueled increased food production. But these nutrients also degrade water quality as they enter inland and coastal waters because runoff can spur harmful algal blooms that threaten human and ecosystem health as well as local economies.
Nutrient inputs to landscapes are poorly monitored in most places. Researchers and land managers rely on models to understand fluxes within watersheds using assumptions derived from broad land use classifications. However, different crops require different types and amounts of fertilizers, and basing models solely on land use can inaccurately represent the locations and quantities of the nutrients entering ecosystems. In addition, these land use classifications miss distinctions within a single land use, like differences between applications of manure and chemical fertilizer in agricultural areas, which have different effects on watersheds.
Hamlin et al. present a novel modeling approach intended to show more accurately where nitrogen and phosphorus originate in the landscape. The resulting tool, the Spatially Explicit Nutrient Source Estimate Map (SENSEmap), quantifies inputs across a swath of the U.S. Great Lakes Basin that includes portions of eight states. With 30-meter resolution, the tool can highlight meaningful patterns at a scale that can inform land use decisions.
Using a technique known as k-means clustering, the study redefines land use categories based on dominant nutrient sources—such as septic systems, agricultural fertilizers, manure, and others—to create what the researchers call nutrient input landscapes (NILs). The nine NILs the team came up with represent agricultural, urban, and rural landscapes and provide more information about nutrient inputs than standard land use classifications. Agricultural landscapes, for instance, are described as either manure dominated, chemical fertilizer dominated, or mixed, whereas urban landscapes are differentiated between septic-served suburbs and sewer-served cities.
With this refined approach, the researchers showed that land use alone does not capture nutrient source dynamics in the Great Lakes Basin. They noted that their model offers insights into landscape variability that are lacking in land use data alone. However, because of uncertainties in the remotely sensed data used in the model, the authors suggested that SENSEmap is most appropriate for management at the watershed scale rather than at smaller field scales. The tool is currently being used within a new decision support system, the Tipping Point Planner, to aid local governments and watershed planners. (Journal of Geophysical Research: Biogeosciences, https://doi.org/10.1029/2019JG005134, 2020)
—Aaron Sidder, Science Writer