Spatial And Temporal Distribution Of Drought And Vegetation Response In Chongqing

In recent years,serious drought disasters frequently occurred in China,which had caused serious losses to agriculture,economy and society.The spatial and temporal distribution of precipitation in humid area with complex terrain and special climate environment is not uniform even if there is sufficient precipitation,which might lead to seasonal drought events.Besides,drought condition recently has been exacerbated with rising air temperatures.Therefore,it is necessary to explore the spatial and temporal distribution of drought disasters in humid areas.As the study area,Chongqing is located in the humid region of China.Based on daily long-term observed meteorological data,the current research calculated existing meteorological drought indices,including relative humidity index?M?,standardized precipitation index?SPI?,standardized precipitation evapotranspiration index?SPEI?,comprehensive drought index?CIspi?and the modified comprehensive drought index?CIwap?.The 5-day soil moisture observations from 2010 to 2013 were used to assess the performaces of these drought indexes for selecting the most suitable meteorological drought index in the study area.Records of long-term meteorological data and soil moisture were obtained from the National Climate Center,China Meteorological Administration?CMA?.By using the optimal meteorological drought index,soil moisture and remote sensing data,the spatial and temporal distribution of drought in Chongqing was explored.Main results are as follows:1.With 5-day soil moistre at 20 cm depth measured at 33 stations from June to September of 2010-2013 as the evaluation criterion,M index was the optimal meteorological drought index compared with SPI,SPEI,CIspi and CIwap at 30-and90-day time scales.The SPI at 30-day time scale,CIspi and CIwap underestimated the drought everity and could not identify some medium drought events.The SPEI at 30-and 90-day time scales not only exaggerated drought severity but also neglected some medium drought events.Moreover,drought classifications identified by M index were in agreement with that of the observed soil moisture.The results of correlation analysis,overall accuracy and kappa coefficient confirmed the accuracy and reliability of M index for monitoring drought in Chongqing.To further verify utility of M index,three stations?Changshou,Tongliang and Zhongxian?were chosed to analyze the relationship between soil moisture,precipitation,potential evapotranspiration,and the optimum drought index in 2011.According to M index,summer drought during 1970 to 2014 happened in each year and showed a slightly upward tendency in recent years,indicating that the increase of evaporation caused by raised air temperature was not ignorable under the global warming condition.2.The optimal meteorological drought index?M index?is usually calculated based on data observed at weather stations,and need to be spatialized for monitoring the spatial distribution of drought.The calculation of M index is based on precipitation and potential evapotranspiration which should be spatialized first.For the estimation of potential evapotranspiration?PET?,14 common models,that are Abtew:Ab,Hamon:Ha,Hargreaves and Samani:HS,Jensen-Haise:JH,Makkink:Ma,Mahringer:Mah,McGuinness-Bordne:MB,Penman:Pe,FAO Penman-Monteith:PM,PenPan:PP,priestley-taylor:PT,Romanenko:Ro,Turc:Tu and WMO were selected and compared based on daily observed meteorological data at 64 stations from 1980 to2009.According to results of precision indexes,the performances of Pe and PP based on radiation-aerodynamics method were obviously better than that of other models.Most of radiation-based models were slightly more accurate than those based on wind speed,and both of them had large errors,which were affected and limited by local topography and climate.The accuracy of the radiation-based models was relatively higher from April to September,while that of the models based on wind speed was relatively lower during the same period.Among these PET models,PM performed best and was used to estimate potential evapotranspiration in the study area.Referring to previous researches,the precipitation and other climatic facors?air temperature,sunshine hours,relative humidity and wind speed?,interpolated by the thin plate spline?TPS?based on cooperation variables?longitude,latitude and elevation?,which were used to estimate potential evapotranspiration with FAO Penman-Monteith?PM?.On this basis,the spatial distribution of the average annual climatic variables and the average annual M index from 1980 to 2009 was analyzed.The results showed that the values of M index in the western,central and part of northeastern Chongqing was relatively lower,where drought was more frequentdue to less prepicitaiton and higher potential evapotranspiration.However,drought frequency in the southeastern and part of northeastern Chongqing was lower with higher M values.3.Based on the observed data of 5-day soil moisture at 20 cm depth from March to September of 2010-2014 and daily air temperature at 33 stations from 2010 to 2014,the performance of different interpolation methods to estimate soil moisture was compared,including the reverse distance weighting method?IDW?,a series of kriging methods,tension regular spline method?RST?and thin plate spline method?TPS?.The results of precision indexes indicated that IDW with simpler calcutation had slightly better performace than other methods and was proved to be the most suitable interpolation method for the soil moisture in the study area.The spatial database of soil moisture from 2010 to 2014 was established by IDW,and the spatial and temporal distribution of drought in the 5 years in the study area was discussed.The results showed that drought occurred every year in the study area,mainly concentrating in August and September.The drought had obvious regional distribution difference and was prone to occur in the western and central Chongqing.The results indirectly comfirmed the feasibility of IDW interpolation for soil moisture,which were consistent with the previous researches.4.Based on 16-day Normalized Differential Vegetation Index?NDVI?,Vegetation Condition Index?VCI?and spatilazed M index,the relationship between drought and vegetation from 2003 to 2009 and the spatial-temporal distribution of drought from June to September in 2006 were analyzed.There were seasonal and spatial differences of distribution of climatic factors and M index.Based on the Pearson correlation coefficient,NDVI and VCI had good consistency and significantly positive correlation with M index,indicating that vegetation grew well with the increase of M index,but vegetation growth was affected with the decrease of M index below the threshold level of drought to some degree.In addition,there was a strong positive correlation between soil moisture and M index in space,but the correlation coefficients decreased with the increase of elevation.Taking 2006 as an example,combined with climatic factors,soil moisture and vegetation growing condition,the spatial and temporal distribution of drought condition from June to September in 2006 was analyzed in this paper.The statistical results of the daily M index showed that the drought began on June 9^th,but the drought condition was not serious.After June 29^th,the drought was dispeared.On July 29^th,the drought appeared in the western Chongqing and began to aggravate and spread gradually.In the middle of August,the drought spread to the central and the northeast of Chongqing.It was not until the beginning of September that the drought condition alleviated.The total number of drouthy days was as high as 52 days,mainly concentrating in the western,central and northeastern Chongqing witha lower elevation.The severe drought and extreme drought were mainly distributed in western Chongqing.The corresponding NDVI and NDVI changing rate(NDVI_var)also had similar changes as M index.During the worst of the drought,the mean NDVI values in areas under medium,severe and extreme drought were 18.67%,23.64%and 31.71%lower than the mean NDVI values in the same region over the same period,respectively.A more detailed analysis on the drought and vegetation conditions during this period based on four typical sites was implemented.The result verified the response of vegetation on drought and further proved the accuracy and applicability of M index for monitoring drought in the study area.In conclusion,the spatial and temporal distribution pattern of drought in the study areaat the daily time scale based on the optimum meteorological drought index,soil moisture,and remote sensing data was explored in current research,which will provide a scientific basis for drought monitoring and evaluation in the study area and be important and significative to the agricultural production and social economy.