Monitoring Change in Floodplain Vegetation Due To River Restoration Treatments With Remote Sensing in the Intermountain West

Abstract

River corridors are essential to life, but have been subjected to centuries of alteration, disconnection, and simplification. Process‐based river restoration has increased in recent years, with the goals of reversing degradation and reestablishing natural processes. Our objectives are twofold: (a) qualify the historic, ecologic, and geomorphic contexts of eight U.S. projects with information gathering, and (b) quantify the causal impact of treatments on remotely‐sensed floodplain vegetation with counterfactual analysis. The most frequent goals were to increase connectivity and support beaver, and the most frequent degradation drivers were grazing/agriculture and beaver removal. There is a >99% probability that treatment increased the floodplain area of vegetation in Oregon at Bridge Creek by an average of 20%, at Whychus Creek by 21%, and at Lost Creek by 29%; and in Idaho at the Yankee Fork by 8%. There is no evidence that treatments affected vegetation area in Idaho at Baugh Creek (>90% probability); or in Colorado at either the South Fork South Platte River (>70% probability) or Fourmile Creek (>70% probability). There is a >90% probability that treatment decreased vegetation area in Idaho at Tincup Creek by 12%. Additional study is needed to determine how key characteristics, such as ecoregion and land use intensity, influence outcomes. We recommend that restoration monitoring pair remote sensing with local knowledge and field data collection to fully measure treatment effects. As ongoing restoration projects mature, our study can serve as a roadmap for further investigation into the long‐term outcomes of river restoration.