Irrigation effects on soil organic carbon differ between annual cropping and crop-pasture rotations in temperate agroecosystems

Given its positive effect on crop production and carbon inputs, irrigation has the potential to increase soil organic carbon (SOC). However, its effect depends strongly on climate and soil type, making SOC responses difficult to anticipate in specific agroecosystems. Our objective was to quantify the effect of irrigation on SOC stocks in Uruguayan agricultural systems, where irrigation is a relatively recent practice. We compared SOC stocks under center-pivot irrigation with those of the paired rainfed corners of the same fields. Under crop-pasture rotations (7 sites), high site-to-site variability prevented a conclusive effect of irrigation on profile SOC (185 ± 12 vs 169 ± 13 Mg ha^-1 at 0-80 cm, P = 0.13), despite a tendency for higher stocks at 0-10 cm. Under continuous annual crops (7 sites), irrigation increased SOC (158 ± 14 vs 143 ± 10 Mg ha^-1, P = 0.02) but, surprisingly, this was explained by differences in intermediate soil layers (20-30 and 30-40 cm, p ≤ 0.01), while no changes were detected in topsoil SOC. Fractionation of SOC in the top 30 cm showed no differences in particulate organic carbon, indicating that SOC responses were related to the mineral-associated organic carbon (MAOC) pool. In crop-pasture rotations, irrigated areas tended to have higher MAOC at 0-10 cm (p = 0.12), whereas under annual crops MAOC increased only at 20-30 cm (p = 0.02). These results suggest that downward transport of dissolved organic compounds and their stabilization as MAOC contribute to SOC accumulation at depth under continuous annual cropping.