Shoaib A. Anees ^1* Kaleem Mehmood ² Sultan Muhammad ³ Khadim Hussain ² MOHAMMAD J. ANSARI ⁴ Sulaiman A. Alharbi ⁵ Fahad Shahzad ² Waseem R. Khan ⁶ Mi Luo ⁷

• ¹ University of Agriculture, Dera Ismail Khan, Dera Ismail Khan, Khyber Pakhtunkhwa, Pakistan
• ² Beijing Forestry University, Beijing, Beijing, China
• ³ University of Swat, Swat, Khyber Pakhtunkhwa, Pakistan
• ⁴ M. J. P. Rohilkhand University, Bareilly, Uttar Pradesh, India
• ⁵ King Saud University, Riyadh, Riyadh, Saudi Arabia
• ⁶ Putra Malaysia University, Selangor Darul Ehsan, Selangor Darul Ehsan, Malaysia
• ⁷ Nanning Normal University, Nanning, Guangx, China

Fractional Vegetation Cover (FVC) is a critical ecological metric that quantifies the proportion of ground covered by vegetation, providing valuable insights into vegetation density and spatial distribution which are essential for environmental monitoring, agricultural management, and biodiversity conservation. Pakistan's diverse ecosystems, including deserts, forests, and wetlands, are highly sensitive to climate change and other environmental challenges. This study used MODIS NDVI data to calculate the nationwide FVC in Pakistan and analyze its spatiotemporal dynamics. Over the past 18 years, Pakistan's FVC has shown an overall increasing trend with an average value of 0.31. The highest value was recorded in 2017 at 0.37, while the lowest value was recorded in 2004 at 0.26, as observed in our data analysis. Trend analyses indicate an increase in FVC from 2013 to 2020, with fluctuations during the period from 2003 to 2012. Due to the widespread distribution of hilly terrain and lower levels of urbanization, FVC in these regions is generally higher, reaching up to 70%. The Hurst exponent, used to measure the long-term memory or persistence in the vegetation dynamics, had an estimated R/S ratio of 0.718. A value greater than 0.5 indicates that the FVC time series data exhibit some degree of long-term memory or autocorrelation. The dynamic drivers of FVC in Pakistan include rainfall, temperature, and the Compounded Night Light Index (CNLI). Through weighted overlay and correlation analysis, it was determined that FVC was positively correlated with rainfall (0.6) and negatively correlated with Land Surface Temperature (LST) (-0.59) and CNLI (-0.43). The high R (0.89) and R2 (0.80) values suggest that the values predicted and observed for the dependent variable are highly correlated. The small RMSE (0.011) value indicates that the model's predictions are reasonably close to the actual values. The variable importance measures from the Random Forest regression model indicate that CNLI (62.40%) is an important predictor, while Rain has the lowest percentage (9.30%). These results suggest that CNLI has the highest importance among different variables. Understanding the interactions between CNLI and FVC is essential for comprehensive monitoring and management of the environment in Pakistan.