One important driver of nutrient loading to lakes is the amount of agricultural land-use cover in its watershed, which is associated with increased nutrient runoff entering lakes. Although broad-scale studies have considered aggregated measurements of agricultural activity such as percentage row-crop or total agricultural land use as drivers of lake water quality, few studies have considered more granular information such as crop type, or the application of manure and fertilizer. Such details may be particularly useful in designing impact assessments, optimizing agricultural practices, and selecting policy instruments to maintain or improve water quality. In this study, we asked: (1) Are more granular measures of agricultural activity related to lake water quality across hundreds of lakes and their watersheds; and, (2) To what extent do lake and watershed characteristics mediate the relationships between agricultural land use and lake water quality? To answer these questions, we modelled lake water quality hierarchically as a function of watershed agricultural activity nested within hydrologic and climatic regions. We compiled agricultural characteristics for over 500 watersheds including both aggregated and granular measures of agricultural activity and combined it with a lake water quality database on thousands of lakes in the midwestern and northeastern US (LAGOS-NE).
Results/Conclusions
Corn and soybean land use cover in our study watersheds ranged from 1 to 63%; whereas watersheds received from 2 to 100 kg/ha and 2 to 75 kg/ha of fertilizer and manure applications respectively. We found that considering detailed agricultural activities such as crop type, manure inputs, and fertilizer application improved our ability to predict lake water quality beyond documenting the percent agricultural land use. Specifically, we found that lake nitrogen concentrations increased with increasing watershed fertilizer use and corn crop cover; whereas, lake phosphorus concentrations increased with increasing manure applications and soybean crop cover. Neither fertilizer use nor manure applications were strongly correlated with total agricultural land use (r < 0.3). However, fertilizer use (r > 0.5) was correlated with row crop land use. Our results highlight the need for future regional to continental-scaled policies and management strategies to consider not only aggregated measures of agricultural activity but also specific agricultural activities relative to lake water quality. Our results also suggest that lake nitrogen and phosphorus concentrations are driven by different agricultural activities and this may affect the outcome of policies to enhance water quality depending on whether they focus on lake phosphorus or nitrogen.