RESUMO
Irrigated maize-based Cropping Systems (CS) are questioned because of the high risk of herbicide transfer to water. An 8-year systemic experiment was conducted to i) compute a multi-performance comparison between a Conventional Maize Monoculture (MMConv) and four CS that aimed to reduce irrigation and herbicide leaching: MMLI, a low-input MM using cover crop and Integrated Weed Management (IWM) techniques; MMStill, a Strip-tillage MM using cover crop; MMCT, a Conservation Tillage MM with cover crop; Maize-MSW, an IWM Maize rotated with Soybean and Wheat and ii) determine the main drivers and evaluate the influence of CS on herbicide leaching in maize. Drainage water was collected through 1-m depth lysimeter plates and analysed for 6 herbicide molecules and 1 degradation metabolite. MMLI yielded 10.7 t ha-1 close to MMConv (11.5 t ha-1) despite a lower herbicide use (-57%) and irrigation (-21%). MMLI and Maize-MSW had less drainage events compared to MMConv. MMCT and MMStill both yielded less (respectively 7.6 t ha-1 and 6.2 t ha-1) while their herbicide use increased (both +24%). Mean annual herbicide losses were 0.5 ± 1.0 g ha-1 for MMLI, 0.7 ± 1.2 g ha-1 for Maize-MSW, 1.3 ± 2.1 g ha-1 for MMStill, 2.0 ± 4.8 g ha-1 MMConv and 3.0 ± 9.6 g ha-1 for MMCT. Herbicide leaching remained variable but was consistently and mostly influenced by drainage volume. According to the CS, only 1.5 to 6.0 drainage events were responsible for 90% of the herbicide losses. High leaching peaks were identified for mesotrione and glyphosate and may indicate that preferential flows occurred, especially under MMCT. Quantity applied had limited influence on herbicide leaching. To reduce the herbicide leaching risk, CS must concomitantly manage water quality and quantity through a combination of agroecological practices, as in MMLI, a CS able to reach other technical objectives. Present study recommends assessing CS through a diversity of performance indicators.
