Analysis of spatiotemporal variation in dissolved organic carbonconcentrations for streams with cropland-dominated watersheds

It remains a challenge to understand how dissolved organic carbon (DOC) is cycled fromfarmlands to rivers due to the complex interaction between farming practices, the baseflow hydrology of predominantly flat lowlands, and seasonal environmental influences such as snowpack. To address this, field DOC concentrations were measured monthly throughout the year at sub-basin scales across the Chippewa River Watershed, which falls within the Corn Belt of the Midwestern United States. These DOC dynamics in stream water from croplands were benchmarked against the data sampled from hilly forested areas in the Connecticut River Watershed. The Soil Water Assessment Tool (SWAT) simulation was applied to provide potential predictive variables associated with daily baseflow. Our study outlines a framework using the combination of primary field data, hydrological modeling, and knowledge-based reclassification of Land Use/Land Cover (LULC) data to analyze the viability of modeling the spatial and temporal variations of cropland stream DOC concentrations. Calibration of the SWAT model resulted in the overall daily Nash-Sutcliffe model efficiency coefficient (NSE) of 0.67 and the corresponding R-2 = 0.89. Our main results show: 1) baseflow DOC concentrations from croplands were substantially higher throughout the year relative to other landcover areas, especially for spring runoff/snowmelt scenarios, 2) an empirical analysis explained similar to 82% of the spatial gradient of annual mean observed DOC concentrations, and 3) with the addition of hydrological simulated variables, a linear model explained similar to 81 % of monthly and 54 % of daily variations of observed DOC concentrations for cropland sub-basins. Our study identified key factors regulating the spatiotemporal DOC concentrations in cropland streamflow; the contribution here promotes to strengthen future analytical models that link watershed characteristics to carbon cycling processes in a large freshwater ecosystem.

Automated summary

This study used field samples, hydrological modeling, and reclassification analysis of land use data to investigate the spatial and temporal variations of dissolved organic carbon (DOC) concentrations in cropland streamflow. The results showed that baseflow DOC concentrations from croplands were significantly higher than those from other landcover areas, especially during spring runoff/snowmelt scenarios. Key factors regulating the spatiotemporal DOC concentrations in cropland streamflow were identified, contributing to the improvement of analytical models linking watershed characteristics to carbon cycling processes.