Investigation of climate, land cover and lake level pattern changes and interactions using remotely sensed data and wavelet analysis

Abstract

Lake Urmia –a hypersaline lake in northwest of Iran- has lost perceptible amount of its waterbody since mid ‘90s. Reasons behind this catastrophe could be categorized as climatic factors and anthropogenic activities. To investigate pattern changes and interactions of the components involved in this dramatic water loss, precipitation (P), discharge (Q), temperature (T), sea surface temperature (SST) alongside satellite images of vegetation cover (VC) from 1990 to 2019 were used for pattern and interaction analysis, including trend and seasonality detection. Owing to the fact that this endorheic lake in extremis is located between East Azerbaijan (EA) and West Azerbaijan (WA) provinces, additionally it was aimed to juxtapose changes in the aforementioned variables assembled from these two provinces. To monitor changes in VC, normalized difference vegetation index (NDVI) time series for two pilot sub-watersheds from EA and WA provinces were calculated using remotely sensed images. Mann-Kendall (MK) trend test revealed that amongst all variables, Q in both sub-watersheds has been intensely decreased in 99% confidence level. Moreover, continuous wavelet transform (CWT) of Q exhibited drastic de-escalations in terms of annual and semi-annual periodicities. VC and urbanization in both sub-watersheds showed significant increasing trends, whereas VC in WA province showed a higher level of trend. Fairly stable P in terms of trend and periodicity alongside intensification in VC concluded intense exploitations of Q which is the other main inflow into this lake. Wavelet coherence transformation (WTC) between all considered variables and Lake Urmia's WL revealed the great impacts of Q, T and SST variables on the lake's WL fluctuations and seasonality. The impact of T and SST on WL fluctuations did not illustrate perceptible changes throughout the study period (i.e., before and after lake's desiccation). The impact of Q on the lake's WL, however, has moved to larger frequency bands with increased phase-lags, and thus, it could be concluded that the intense reduction of the lake's WL is in proportion to intense reduction of rivers discharge to this lake; leaving transparent traces of anthropogenic impacts on Lake Urmia's tragedy.