Introduction: Drought is a complex climatic phenomenon with severe impacts on agriculture, water resources, and food security (OECD, 2025). Agricultural drought, as one of the most important drought types, is most closely related to food security because reduced soil moisture limits plant growth and crop production (Wang & Ren, 2025). Remote sensing, as a newly emerged technology, overcomes the spatial limitations of traditional monitoring, enabling timely, large-scale assessments and on-time, precise management practices. It also helps researchers predict the future state of agroecosystems, which is important for policymakers.The Vegetation Health Index (VHI), which integrates vegetation and temperature conditions, is one of the most effective indicators for agricultural drought monitoring (Kogan, 1995). Climate change intensifies droughts by altering rainfall patterns and increasing temperatures (IPCC, 2021; Dikshit et al., 2022). This study projects future temperature, precipitation, and VHI trends in Khorasan Razavi Province involving CMIP6 scenarios using Google Earth Engine analysis.

Materials and Methods: Khorasan Razavi Province in northeastern Iran, characterized by an arid to semi-arid climate, was selected as the study area. MODIS satellite data (MOD13Q1 for NDVI and MOD11A2 for LST) were used to calculate the Vegetation Health Index (VHI = 0.5×VCI + 0.5×TCI) for monitoring agricultural drought for 2023. Future climate projections were obtained from five CMIP6 GCMs (CanESM5, BCC-CSM2-MR, ACCESS-CM2, IPSL-CM6A-LR, HadGEM3-GC31-MM) under SSP2-4.5 and SSP5-8.5 scenarios. Daily temperature and precipitation data from 10 meteorological stations (1994–2013) were used to validate model performance using R², Root mean Square Error (RMSE), and Mean Bias Error (MBE). The two best-performing models were selected for future simulations. VHI for near-future (2045–2060) and far-future (2085–2100) periods was computed. All analyses and map generation were conducted in Google Earth Engine and ArcMap 10.8.2, enabling high-resolution spatial and temporal assessment of agricultural drought trends.

Results and Discussion: Statistical metrics (RMSE, MBE, R²) indicated varying performance among the CMIP6 models; IPSL-CM6A-LR was most suitable for maximum temperature and precipitation, while CanESM5 performed best for minimum temperature. Near-future projections (2045–2060) indicate maximum temperature increases of up to 0.7-3 °C under SSP2-4.5 and up to 5 °C under SSP5-8.5. In the far future (2085–2100), warming intensifies, with increases up to 7 °C. Minimum temperature also rises steadily, with stronger changes under SSP5-8.5. Precipitation shows high variability, with overall reductions, especially in spring months. Near-future decreases reach 14–15 mm in some months, while far-future declines are more severe. Spatially, the southern and eastern parts of the study area are most vulnerable to drought. VHI values remain below the drought threshold (<40) throughout the period 2045–2100. SSP5-8.5 shows a sharper long-term decline in VHI, despite a short-term CO₂ fertilization effect. County-level analysis highlights Salehabad, Torbat-e Jam, Bakharz, Taybad, Khaf, and Zaveh as the most affected areas.

Conclusion: Future climate change will intensify warming, reduce precipitation, and worsen drought conditions in Khorasan Razavi which would affect agroecosystems directly or indirectly. These results emphasize the need to revise cropping patterns, manage water resources, and implement climate change adaptation policies in Khorasan Razavi Province, especially in areas that are expected to be most affected.

Keywords: Agricultural drought; Climate change; Climate scenarios; Genral circulation models; Google Earth Engine