Research On Drought Monitoring Methods Taking Into Account The Impact Of Climatic Indices

Drought is one of the world's most severe natural disasters,which is characterized by high frequency,long duration and large impact area.The occurrence of drought has an impact on agricultural production,ecology and socio-economics.It causes different degrees of loss in the global scope.Climate anomalies caused by global climate change and sea-air interactions increase the likelihood of extreme events.Further analysis and understanding of the remote-related mechanisms of drought occurrence and the construction of integrated drought monitoring models that take into account climatic anomalies are the main ways to prevent and reduce the impact of droughts and reduce losses.In China,which has a vast territory,the mechanisms by which drought is influenced by the ENSO and MJO vary within different eco-geographical regions.Through the study of the teleconnection of the ENSO and MJO with drought in different regions,the mechanism of drought occurrence can be further understood.On the basis of this analysis,a comprehensive drought monitoring model was established based on a machine learning random forest algorithm by integrating the meteorological drought index,remote sensing drought index,land cover and other factors reflecting regional differences in drought.The main research methods and conclusions are as follows.:(1)Meteorological drought indices(PDSI/SC?PDSI/SPEI of 1,3,6,9,and 12 months scales)from 1982 to 2018 were constructed based on the meteorological data of 577 stations in China,and the remote sensing drought indices(VCI/TCI)were obtained for the 2006–2019.The spatial and temporal correlations between climate anomalies and drought events in48 eco-geographic regions of China were analyzed to obtain the characteristics and spatial and temporal variations of the effects of climate anomalies on drought.The results show that,in general,climate anomalies have a strong correlation with drought,and the spatial correlation differs in different ecological regions.When ENSO cold and warm events occur or MJO oscillations are abnormally high or low,the response of drought to extreme events varies in different regions.The impact of climate anomalies on agricultural drought was higher than that of meteorological drought,with the highest correlation coefficient of 0.72 in regions such as the Qinghai-Tibet Plateau and Xinjiang,while the impact of ENSO and MJO on meteorological drought was significant in Jiangnan with the highest correlation coefficient of0.40.Compared with meteorological drought,the response of agricultural drought to climate anomalies also showed a significant lag,with a lag time of 6 months,and the correlation between the two was most pronounced and highest,with a mean correlation coefficient of0.38 and a maximum of 0.75.(2)In this paper,the spatial and temporal correlation of MJO and ENSO with drought in different eco-geographical regions is considered,and both of them and the eco-geographical regions are included in the independent variables of the model construction.Other independent variables include the remote sensing drought indices VCI and TCI,the meteorological drought index SPEI,and 11 eigenvalues of land cover,elevation,aspect,slope and available water content,while the dependent variable is PDSI.The model was divided into four sub-models based on seasonal differences in drought,and its accuracy was verified.Through simulation,the correlation coefficients between the results of the four models and the real values reached above 0.95 with R~2 around 0.90 for the whole data set,and above 0.83with R~2around 0.70 for the test set.(3)Based on the constructed model,drought monitoring was simulated for a total of 156months from 2006-2018 in North China and the Yellow Huaihai Sea region.And six validation stations were selected to compare the interpolation results of PDSI,meteorological drought index and precipitation data from 2006-2018 to further validate the drought monitoring capability of the newly constructed model.The model has good performance and is consistent with the actual drought conditions.From the station time series,the trend of monitoring results has a consistent trend change correspondence with precipitation,which can reflect the drought situation well.(4)Typical drought events in the drought-prone regions of Anhui,Yunnan and Liaoning in China were selected to further validate the monitoring results of the new model and to evaluate the accuracy of the model,and the results showed that the model monitoring results were generally consistent with the actual situation.This study integrates data from multiple sources,and the drought monitoring model constructed to take into account the effects of abnormal climate change can be applied to drought monitoring in North China and the Yellow and Huaihai Seas,Anhui,Liaoning and Yunnan.The quantitative monitoring in space is achieved,providing new ideas and tools for drought resilience in China.