Detecting the effects of land use/land cover on mean annual streamflow in the Upper Mississippi River Basin, USA

- We model impact of LULC and other factors on regional mean annual streamflow.
- LULC impact is significant compared with other climatic and geomorphologic factors.
- LULC impact is not a simple function of spatial extent.
- Different measures might show different patterns/magnitudes of LULC impact.

SummaryThe effects of land use/land cover (LULC) along with other climate and geomorphologic factors on mean annual streamflow in the Upper Mississippi River Basin (UMRB) were explored using a nonlinear model applied to a spatial dataset of more than 180,000 catchments. The model was linearized and solved via a geospatial regression model to deal with spatial dependency in data. Results show that LULC is a very important agent with respect to its impact on mean annual streamflow in UMRB. Compared with other climate and geomorphologic factors, the signature of LULC on streamflow which was shown through various measures in this study is evident and significant. The magnitude of the impact on streamflow varies from one LULC to another. It is not a simple function of a LULC's spatial extent but arguably a result of complex interactions among various LULCs as well as other climate and geomorphologic factors. Our analysis indicates that caution needs to be taken in comparing different studies or in generalization across scales regarding the impact of LULC on streamflow. It is because the result of a study is not only the outcome of the geophysical processes observed at a particular spatial and temporal scales but also a product of the approach, model, variables, and/or measures used in the study. Methodologically the proposed model provided an effective way to utilize an extensive spatial dataset of various climatic, geomorphologic, and LULC variables for a large region like UMRB to assess and compare the impact of various factors on mean annual streamflow at regional scale. Furthermore, the model was able to handle spatial dependency in data while avoiding the common problem of nested input in hydrologic modeling.