Investigating the effect of temperature, precipitation, and soil salinity changes on Riparian Forests’ phenology using a remote sensing approach

Abstract

Many previous studies have shown the sensitivity of the plant's phenological cycle to the changes of temperature and precipitation. Other parameters such as pests, salinity and water stresses during the growing season also have a significant effect on changes in the phenological cycle of vegetation. Due to the importance of plants' phenological cycles, this study focuses on the effect of long-term stresses caused by temperature, precipitation and salinity of water and soil on the change of phenological cycle of riparian forest ecosystem around Karun River, Iran. Riparian forests are one of the most important forest ecosystems in arid and semi-arid regions. The dataset includes MODIS products and ETM+, OLI Landsat images of the study area from 2000 to 2020. NDVI, EVI and salinity indices SI1, SI2, SI3 were computed from spectral bands. The results showed that the start of growing season (SOS) with a decreasing slope of 0.02–2.4, the peak of the growing season with a slope of 1.5–4.5 and the end of growing season (EOS) with a slope of 12–22 days per year are subjected to change over the observed years. These changes have shortened the length of growing season (LOS) of plants in 2011 and 2012. The results also show that precipitation and temperature changes cannot fully explain the shift at the SOS, peak and EOS of the growing season. However, the coincidence of soil and water salinity increase in the region -resulted by the dewatering of Gotvand Dam (in 2011) with shorter LOS of riparian forest in the period 2020–2010 has been observed. In addition, the phenology cycle based on NDVI and EVI of both sensors indicates a shorter growth season in 2011–2020 in comparison to its earlier decade. Landsat-based vegetation indices are more accurate in extracting the Riparian forests' phenological information compared to MODIS-based indices. Although the phenology cycles based on Landsat series were more detailed and closer to the field observation compared to the MODIS medium resolution images, MODIS products provide acceptable results especially in monitoring the changes of timing the phenological events in different years.