||Understanding the spatially and temporally variant phenological responses and cycles can greatly assist the administrative planning, policy making and management in grazing, planting, and ecosystem conservation. The linkages of analysis as a basis for management have received increasing attention in the context of climate change. This research focuses on analyzing phenological responses of vegetation as constrained and moderated by environmental factors, such as landscape and season, in the geographically diverse Upper Colorado River Basin (UCRB). Due to the geographic diversity of phenological forcing in the UCRB, several homogeneous phenological subregions (phenoregions) are delineated, and the phenological responses of vegetation are analyzed on a per phenoregion basis. A multivariate adaptive regression splines (MARS) approach is adopted to model and interpret the regionally and seasonally specific relationships between environmental drivers (temperature, precipitation and solar radiant energy) and vegetation abundance, indicated by a Vegetation Index (VI). Short-term predictions of vegetation abundance are made using the models. Taking into consideration the scale of the study area and the time-step of the models, 1 km 7-day interval eMODIS data and the 1 km NASA AMES Ecocast data are used to articulate the dependent and independent variables. The series of models are integrated into a prototype phenological Decision Support System (DSS) to provide predicted vegetation abundance over the growing season and the trends of climatic variables leading to potential grazing management strategies. The implementation of the DSS is a unique attempt to integrate phenological theory and GIS technology, the combination of which makes this DSS analytically-based, intuitive and more user-friendly.