Most studies assessing vegetation response following control of invasive Tamarix trees along southwestern U.S. rivers have been small in scale (e.g., river reach), or at a regional scale but with poor spatial‐temporal replication, and most have not included testing the effects of a now widely used biological control. We monitored plant composition following Tamarix control along hydrologic, soil, and climatic gradients in 244 treated and 172 reference sites across six U.S. states. This represents the largest comprehensive assessment to date on the vegetation response to the four most common Tamarix control treatments. Biocontrol by a defoliating beetle (treatment 1) reduced the abundance of Tamarix less than active removal by mechanically using hand and chain‐saws (2), heavy machinery (3) or burning (4). Tamarix abundance also decreased with lower temperatures, higher precipitation, and follow‐up treatments for Tamarix resprouting. Native cover generally increased over time in active Tamarix removal sites, however, the increases observed were small and was not consistently increased by active revegetation. Overall, native cover was correlated to permanent stream flow, lower grazing pressure, lower soil salinity and temperatures, and higher precipitation. Species diversity also increased where Tamarix was removed. However, Tamarix treatments, especially those generating the highest disturbance (burning and heavy machinery), also often promoted secondary invasions of exotic forbs. The abundance of hydrophytic species was much lower in treated than in reference sites, suggesting that management of southwestern U.S. rivers has focused too much on weed control, overlooking restoration of fluvial processes that provide habitat for hydrophytic and floodplain vegetation. These results can help inform future management of Tamarix‐infested rivers to restore hydrogeomorphic processes, increase native biodiversity and reduce abundance of noxious species.

Arable weeds are key organisms for biodiversity maintenance and ecosystem service provision in agroecosystems. Disentangling the drivers of weed diversity is critical to counteract the global decline of farmland biodiversity. Even if distinct scale-dependent processes were alternatively proposed, no general framework unifying the multi-scale drivers of weed dynamics has yet emerged. Here, we investigate the joint effects of field- and landscape-scale processes on weed assemblages in 444 arable fields. First, field margins sheltered greater weed diversity than field core, evidencing their role as biodiversity refugia. Second, community similarity between field core and margin decreased with the distance to margin, highlighting a major role of local dispersal. Third, weed diversity at field margins increased with organic field cover in the landscape, pointing out massive regional dispersal. Fourth, while both local and landscape dispersal explained up to 41% of field core weed diversity, crop type strongly modulated their strength, depicting an intense filtering effect by agricultural management. This study sheds new light on the complex multi-scale interactions shaping weed diversity, field margins playing a key role by strengthening regional dispersal and sustaining local dispersal. Land-sharing strategies improving habitat heterogeneity both locally and regionally should largely promote agroecosystem multifunctionality and sustainability.

Several governmental programs have been established throughout Canada to foster agriculture sustainability. As a best management practice, cover crops (CCs) limit soil erosion and prevent nutrient losses in agroecosystems. Yet, the variable effects of CCs on cash crop productivity previously reported may limit their large-scale adoption by farmers. To address this variability, we conducted an unweighted meta-analysis including 2274 observations from 86 field studies conducted under humid temperate climate to evaluate yield response to CCs for three annual cash crops. Overall, CCs increased corn and small grain cereal yields by 13% and 22% respectively, but did not affect soybean yield. Legume CCs alone or mixed with grasses provided the highest small grain cereal and corn yield increases compared with non-legume broadleaf and grass CCs. CC benefits increased with nitrogen (N) content in CC aboveground biomass but decreased when N fertilizer inputs applied to corn exceeded 60 kg N ha⁻¹. Greater precipitation and N fertilizer inputs reduced the negative effect of grass CCs on corn yield, while benefits of legume CCs were highly resilient to precipitation variations. CC benefits on corn yield increased through time and at low soil organic matter content, especially at low N fertilizer inputs. These results evidence the complex interplay between cash crop productivity, CC management, and environmental factors — related to N inputs from CCs, changes in soil properties (e.g., increased organic matter, improved soil structure or microbial activity), or potential competition for water under drier conditions — which provide new perspectives to promote CC inclusion in cropping systems.