Climate-induced shifts and lengthening of growing seasons across Pakistan’s major vegetation types

Abstract

Pakistan’s diverse ecology, ranging from arid to semi-arid zones and mountainous regions, makes vegetation phenology highly sensitive to climate variability. However, a long-term assessment of phenological shifts across its major ecological zones remains limited at national-scale. This study provides the first consistent nationwide analysis of changes in vegetation phenology metrics from 2001 to 2023, including the start of season (SOS), length of season (LOS), and end of season (EOS). The assessment focused on natural and semi-natural vegetation classes, excluding croplands, to understand climate-driven phenological responses. The MODIS phenology product was used to extract phenological indicators, while vegetation land cover types were derived from the MODIS land cover type product. The impact of climate variability on phenology was assessed using temperature and solar radiation data from the ECMWF ERA5 (European Centre for Medium-Range Weather Forecasts Reanalysis 5) and precipitation data from CHIRPS (Climate Hazards Group InfraRed Precipitation with Station data). Temporal trends in phenological events and climatic variables were analyzed at seasonal and annual scales, with correlations assessed between each climatic driver and phenological parameters. Results indicated variation in SOS (May to August) and EOS (November to March), reflecting phenological differences across different vegetation types across the biomes. A significant trend of delayed SOS (0.10 days/year) and EOS (0.40 days/year) was observed. SOS onset for grassland and savannas were negatively correlated with temperature (−0.424 and −0.422), and consistently indicated negative association with solar radiation across all vegetation types, suggesting an earlier SOS with warmer temperature and bright sunny conditions. Among the divergent vegetation classes (mixed forests, open and close shrublands, and savannas and woody savannas), within-season precipitation showed positive correlations with EOS and LOS (r = 0.455–0.727), indicating moisture-constrained growing season length. Soil moisture exhibited stronger associations with EOS and LOS, particularly in mixed forests (EOS r = − 0.557, LOS: r = −0.71) with higher significance than with SOS, suggesting a greater role in regulating senescence and growing season duration than green-up onset. These asymmetric phenological responses to different climate controls suggest the complex influence of climatic conditions at different stages of vegetation growth, varying across vegetation types and landscapes. The reliance on the MODIS phenology product may introduce uncertainties, suggesting further validation and independent phenology observation across the landscape.