Predicting the partitioning between aqueous and gaseous C across landscapes is difficult because many factors interact to control CO2 concentrations and removal as DIC. For example, carbonate minerals may buffer soil pH so that CO2 dissolves in porewaters, but nitrification of fertilizers may decrease pH so that carbonate weathering results in a gaseous CO2 efflux. Here, we investigate CO2 production and dissolution in an agricultural, first-order, mixed-lithology humid, temperate watershed. We quantified soil mineralogy and measured porewater chemistry, soil moisture, and pCO2 and pO2 as a function of depth at three hillslope positions for a year. The variation of soil moisture along the hillslope was the dominant control on the concentration of soil CO2, but mineralogy acted as a secondary control on the partitioning of CO2 between the gaseous and aqueous phases. The regression slopes of pCO2 vs. pO2 in the carbonate-bearing soils indicate a deficit of CO2 relative to O2 (p < 0.05). Additionally, we found no abiotic gaseous CO2 efflux from carbonate weathering. We concluded that in the calcareous soils, about a third of respired C dissolves and drains from the soil rather than diffusing out to the atmosphere. To represent the global scope of the reactions we evaluated at our local watershed, we used databases of carbonate minerals and land uses to map types of soil degassing behaviors. Based on our maps, the partitioning of respired soil CO2 to the aqueous phase may be globally common and should be accounted for in ecosystem C budgets and models.

An agricultural watershed, Cole Farm, was established as the newest of the three subcatchments in the Susquehanna Shale Hills Critical Zone Observatory (SSHCZO) in 2017. The catchment contains mostly pasture and crops, with a small portion of deciduous forest. The observations in Cole Farm afford an opportunity to test the spatially distributed land surface hydrologic model, Flux-PIHM, in farmland for the first time. In this study, we calibrated the model to only discharge and groundwater level observations at Cole Farm, but it’s able to capture the variations and magnitudes of soil moisture, latent heat (LE) and sensible heat (H) fluxes. Modeled soil moisture on the ridge top matched the observations well, but modeled soil moisture in the mid-slope differed from observations likely due to the existence of fragipan in the soil column. Flux-PIHM reproduced the seasonality and diurnal variations of watershed-average evapotranspiration (ET), sensible heat flux (H), though modeled ET in summer is about 25% greater than tower ET. To study the impact of land cover on hydrology, we imposed two different LAI forcings to the model: spatially distributed versus uniform LAI. Spatially distributed LAI produced higher ET and lower soil moisture in the forested part of the watershed due to higher LAI of deciduous forest in comparison to crops and pasture. But the impact of different LAI forcings on discharge was small. We further compared the water budget simulated by Flux-PIHM in the agricultural watershed (Cole Farm) to a forested watershed (Shale Hills). Flux-PIHM simulated less discharge and higher transpiration and bare soil evaporation in the Cole Farm watershed relative to Shale Hills watershed. Our work shows that with a few key observations, Flux-PIHM can be calibrated to simulate agricultural watershed hydrology, but spatially distributed LAI and soils data are needed to capture the spatial variations in soil moisture and ET.