Coherency and phase delay analyses between land cover and climate across Italy via the least-squares wavelet software

Abstract

Land cover and climate monitoring is a crucial task in agriculture, forestry, hazard management, and ecosystems assessment. In this paper, normalized difference vegetation index (NDVI), land surface temperature (LST), and land cover products by the moderate resolution imaging spectroradiometer (MODIS) as well as precipitation were utilized to monitor the spatiotemporal dynamics of vegetation and climate along with their correlation and coherency across Italy during 2000–2021. The analyses were performed on both pixel and ecoregion levels via the least-squares wavelet software (LSWAVE). It was found that relatively more areas in all ecoregions had positive NDVI gradients than negative for each month since 2000. It was estimated that the average NDVI has increased by ∼0.07 since 2000 for all ecoregions. Except the southern ecoregion which showed an insignificant daytime cooling, other ecoregions have been warming by less than 0.05 °C/year since 2000. Furthermore, precipitation had an insignificant decreasing trend for almost all ecoregions over the past two decades. The annual coherency between NDVI and LST was found much stronger than the annual coherency between NDVI and precipitation. The annual cycles of NDVI and LST were out-of-phase for the southern ecoregion while the annual cycle of precipitation led the one in NDVI by about one month for this ecoregion, the only ecoregion showing the highest Pearson correlation (53%) and annual coherency (39%) between NDVI and precipitation. For other ecoregions, the annual cycles of NDVI and LST were approximately in-phase, i.e., less than a month phase delay.

Keywords

• Coherency
• Ecoregion
• Land cover
• LST
• LSWAVE
• MODIS