Citation:

Dietrich, M., H. Golden, J. Christensen, C. Lane, AND M. Dumelle. Lake chlorophyll-a linked to upstream nutrients across the CONUS. American Geophysical Union (AGU) Meeting 2024, Washington, DC, December 09 - 13, 2024.

Impact/Purpose:

Nitrogen (N) and Phosphorus (P) are known to cause or exacerbate harmful algal blooms in the nation's lakes. This abstract explicitly looks at if river N&P can be linked to lake N&P and chlorophyll-a. The statistical models are based on existing USGS data and the results of the model show the linkage as well as identifying the relevant predictor variables in a parsimonious model. 

Description:

Excess nutrients in the environment can lead to adverse outcomes for humans and ecosystems. Nutrients, often sourced from agricultural-rich areas, may end up in downstream water bodies such as lakes or oceans, leading to eutrophication. We conducted a thorough, data-rich analysis of chlorophyll-a (Chl-a) concentrations in lakes and total nitrogen (TN) and total phosphorus (TP) concentrations in rivers across the conterminous United States (CONUS) to assess watershed scale connections between the two. Chl-a is used here as a proxy for algal biomass, which can be indicative of harmful algal blooms and is commonly correlated to nutrient concentrations within waterbodies. However, there has been minimal research examining the linkages of upstream riverine TN and TP to downstream lake Chl-a at large watershed scales and across disparate climatic and physiographic regions. Our CONUS-wide analysis revealed a significant positive relationship between measured TN and TP concentrations in upstream rivers and Chl-a concentrations in downstream lakes at the watershed scale. Furthermore, we applied random forest machine learning and found that only a small number of explanatory variables (2-3 per model) are needed to accurately predict (71%-83% accuracy) classifications of high or low riverine TN, TP, or lake Chl-a concentrations throughout the CONUS at the watershed scale. Predictor variables of vegetation type, runoff, tile drainage, temperature, and nitrogen inputs were most important in the models, revealing that landscape processes are likely driving most of the variance in nutrient and Chl-a concentrations. Our work supports the notion that rivers act as a conveyor system, supplying nutrients that enhance Chl-a concentrations in lakes within downstream watersheds and demonstrates the power of parsimonious models to elucidate primary landscape factors related to nutrient concentrations and eutrophication across the CONUS.

Record Details: