Agricultural production systems in New Mexico (NM) are under increased pressure due to climate change, drought, increased temperature, and variable precipitation, which can affect crop yields, feeds, and livestock grazing. Developing more sustainable production systems requires long-term measurements and assessment of climate change impacts on yields, especially over such a vulnerable region. Providing accurate yield predictions plays a key role in addressing a critical sustainability gap. The goal of this study is the development of effective crop yield predictions to allow for a better-informed cropland management and future production potential, and to develop climate-smart adaptation strategies for increased food security. The objectives were to (1) identify the most important climate variables that significantly influence and can be used to effectively predict yield, (2) evaluate the advantage of using remotely sensed data alone and in combination with climate variables for yield prediction, and (3) determine the significance of using short compared to long historical data records for yield prediction. This study focused on yield prediction for corn, sorghum, alfalfa, and wheat using climate and remotely sensed data for the 1920–2019 period. The results indicated that the use of normalized difference vegetation index (NDVI) alone is less accurate in predicting crop yields. The combination of climate and NDVI variables provided better predictions compared to the use of NDVI only to predict wheat, sorghum, and corn yields. However, the use of a climate only model performed better in predicting alfalfa yield. Yield predictions can be more accurate with the use of shorter data periods that are based on region-specific trends. The identification of the most important climate variables and accurate yield prediction pertaining to New Mexico’s agricultural systems can aid the state in developing climate change mitigation and adaptation strategies to enhance the sustainability of these systems.
A multiple-frame approach to crop yield estimation from satellite- remotely sensed data
