Depressed attainable wheat yields under continuous annual no-till agriculture suggest declining soil productivity

Abstract

The increase in annually cropped area that has occurred in the eastern Pampas of South America (Uruguay) since 2002 was largely achieved by converting crop-pasture rotations to continuous annual no-till cropping systems. We studied the impact of this intensification on the attainable wheat yield (Yatt) and the yield gap (Yg), with the latter defined as the difference between Yatt and actual yield (Ya). Our measure of agricultural intensification is the length in years of continuous cropping (YCC). We computed both Yatt and Yg from 1072 wheat yield records obtained from producers between 2009 and 2012. The database included grain yield, agro-climatic region, preceding crop, sowing date, cultivar, nitrogen (N) fertilization rate and YCC. For each field, we calculated a climatological index (CI) that combines temperature, radiation and precipitation during both the spike and grain growth phases. To estimate Yatt, we fitted a stochastic frontier production function that uses CI, cultivar, region and YCC as explanatory variables, and includes a term representing the inefficiency function. The latter quantifies management practice effects on Yg. A relative yield gap (RYg) was calculated as the ratio Yg/Yatt. We used regression trees to uncover relationships between RYg and variables integrated in the inefficiency function. The mean Yatt was 4.9±0.7Mgha−1. Its variation was a function of the cultivar and the interaction of CI and YCC. The interaction term indicated that Yatt was highest and independent of YCC when the CI was favorable for wheat growth (high CI, sunny and temperate before flowering and during grain filling as opposed to low CI, rainy and warm), but declined rapidly with YCC when CI was unfavorable for wheat growth. The average Yg and RYg were 1.5±0.8Mgha−1 and 0.31±0.17. The fertilizer N rate explained most of the variation in RYg. When the CI was low, YCC was the second most important variable explaining RYg. When the CI was high, the variation in RYg was best explained by the preceding winter crop. Our results indicate that under low CI, intensifying crop production with continuous annual no-till cropping can cause a measurable reduction in Yatt and Ya. The negative effect of YCC on Ya was significantly higher under weather conditions unfavorable for wheat growth, increasing the RYg. These effects could not be compensated for with the maximum N fertilization rates used by producers under low CI or with a rotation of annual crops. Our interpretation is that shifting cropping systems from a crop-pasture rotation to continuous annual no-till cropping generates a progressive limitation in soil productivity that reduces wheat yield and increases its inter annual variability.