Ge Song,Jin-ting Hung,*,Bo-hn Ning,Ji-wei Wng,Lei Zeng

a Xi’an University of Science and Technology, Xi’an 710054, China

b Xi’an Center, China Geological Survey, Xi’an 710054, China

Keywords:NDVI Groundwater level Climate change Arid land Hydrogeological survey engineering Tarim Basin Qaidam Basin Western China

ABSTRACT At present,investigation about the relationship between the change of groundwater level and vegetation mostly focuses on specific watersheds,i.e.limited in river catchment scale.Understanding the change of groundwater level on vegetation in the basin or large scale,be urgently needed.To fill this gap,two typical arid areas in the west of China (Tarim Basin and Qaidam Basin) were chosen the a typical research area.The vegetation status was evaluated via normalization difference vegetation index (NDVI) from 2000 to 2016,sourced from MODN1F dataset.The data used to reflect climate change were download from CMDSC (http://data.cma.cn).Groundwater level data was collected from monitor wells.Then,the relationship of vegetation and climate change was established with univariate linear regression and correlation analysis approach.Results show that: Generally,NDVI value in the study area decreased before 2004 then increased in the research period.Severe degradation was observed in the center of the basin.The area with an NDVI value > 0.5 decreased from 12% to 6% between 2000 and 2004.From 2004 to 2014,the vegetation in the study area was gradually restored.The whole coverage of Qaidam Basin was low.And the NDVI around East Taigener salt-lake degraded significantly,from 0.596 to 0.005,2014 and 2016,respectively.The fluctuation of groundwater level is the main reason for the change of surface vegetation coverage during the vegetation degradation in the basin.However,the average annual precipitation in the study area is low,which is not enough to have a significant impact on vegetation growth.The annual average precipitation showed an increase trend during the vegetation restoration in the basin,which alleviates the water shortage of vegetation growth in the region.Meanwhile,the dependence of surface vegetation on groundwater is obviously weakened with the correlation index is −0.248.The research results are of some significance to eco-environment protection in the arid area of western China.

1.Introduction

Vegetation plays an important role in maintaining the carbon cycle and regulating climate.Water resources shortage and fragile eco-environment are the major issue in the arid area.Owing to the low petition of water resources for human life demand and vegetation water use,natural precipitation cannot meet the regular growth of plants (Li XY et al.,2002;Zhao WZ and Liu H,2006).Therefore,groundwater becomes the main source for vegetation to survive in the arid area(Rodriguez I,2000; Guo ZR and Liu HT,2005; Fan ZL et al.,2008).In recent years,global warming and human exploitation activities aggravated caused vegetation degradation and then desertification more seriously.In recent years,global warming and human exploitation activities aggravated caused vegetation degradation and then desertification more seriously (Scanlonbe BR et al.,2006).Under this premise,investigation the effect of groundwater level change on vegetation in the arid area of western China has important guiding significance for the rational utilization of regional groundwater,water resources management,and the promotion of eco-environment construction in western China.

Researches on the relationship of groundwater and vegetation mainly focus on the response mechanism of vegetation to groundwater and the relationship between water level fluctuation and plant distribution pattern.For example,Rains MC et al.(2004) established a groundwater and vegetation correlation model for reservoir operation and simulated different situations.Results show that the fluctuation of shallow groundwater level caused by reservoir operation has a significant impact on the distribution of groundwater-dependent plants.Orellana F et al.(2012)analyzed different methods of a field survey on vegetation water use and summarized the existing interaction model between groundwater level and plants.Moffett KB et al.(2012) proposed the concept of the concept-“ecohydrological zone” by establishing a quantitative analysis of salt marsh area,reflecting the combined effects of topography,sediment,and vegetation heterogeneity.On the relationship of vegetation and groundwater in the wet and arid areas of western China,the existing research results are also plentiful.Song CY et al.(2008) using bi-directional indicator species analysis and detrended canonical correspondence analysis to evaluate the distribution and influencing factors of typical plant communities in the Yellow River Delta,which confirms that the distribution of typical plant communities reflected the changes in groundwater depth and soil salinity which determined by geomorphology.Ma YL et al.(2013) studied the typical vegetation in the Yellow River Delta and found that groundwater table depth significantly affected soil salinity,and then affected the distribution of natural vegetation.Zhao CY et al.(2008) and Zhao CY et al.,2003 studied the coupling relationship between vegetation change and groundwater movement in Sangong River Basin and Heihe River Basin.They discovered the vegetation coverage in the basins is restricted by groundwater level,and the groundwater depth determines the vegetation type in the basins.

Tarim Basin is the most environmentally vulnerable area with a high incidence area of land desertification in China.In recent years,research on the interaction between groundwater level and vegetation change in Tarim Basin mainly focuses on the lower reaches of Tarim River,and mainly shadows on the response of dominant communities to groundwater level fluctuation (Wang XY et al.,2017; Chen YN et al.,2003; Bai YF et al.,2016; Yang PN,2005; Xu HL et al.,2007; Ma XD et al.,2011; Wang XY et al.,2016; Hao XM et al.,2008).Therefore,based on existing research,it is necessary to develop larger-scale research.In this paper,Tarim Basin and Qaidam Basin are chosen as the typical research area.Remote sensing data and groundwater level monitoring data are analyzed by linear regression analysis and correlation analysis.The research results can enrich the achievements on the effects of groundwater level on vegetation in the arid area at a large spatial scale.Furthermore,exploring the relationship between groundwater change and vegetation in Tarim Basin is of great significance to maintaining species diversity and ecological balance in the study area.

2.Overview of the study area

2.1.Topography and hydrogeological condition

The Tarim Basin is located in the south of Xinjiang,with an area of 400000 km2.It is the largest arid inland basin in China.The basin as a whole is the pendular distribution with the gravel Gobi edge.Taklimakan Desert is located in the center of the Tarim Basin with an area of about 335000 km2.Among the basin edge and desert,landforms are alluvial fans,alluvial plains,and oases.There are many river systems in the basin,Yarkand River,Aksu River,Kaidu-Kongque River,Weigan River,Kashgar River,Qarqan River,etc.Most of the rivers are sources from a mountainous area,which are recharged by ice-snow melting water in the alpine region,forest precipitation in the middle mountain belt,and bedrock fissure water in the low mountain belt.Total surface runoff in the study area is about 398 × 108 m2(Chen YN,2014).Regional water resources mainly come from the precipitation of blocking wet airflow in mountainous areas around the basin.The piedmont alluvial-diluvial plain of the basin becomes the recharge-runoff area of groundwater.Distribution of groundwater depth in the basin with an annular band.

Qaidam Basin is located in the north of Qinghai-Tibet Plateau,with an area of about 27.6 × 104km2and an elevation of 2652 m to 6600 m.The basin is surrounded by mountains,with the terrain slope gradually decreases from the piedmont to the central plain.Aquifer structure in basin presents zonal distribution from mount front to basin center.The lithology of the aquifer changes from coarse to fine,and the permeability changes from strong to weak.There are 64 lakes in the basin,48 of those are salt lakes.

2.2.Climate and ecological environment

The temperature zone in the region includes the middle temperate zone and the warm temperate zone.The average annual temperature in the middle temperate zone is lower than 10°C,and in the warm temperate zone is higher than 10°C.Precipitation in the basin mainly comes from westerlies with annual precipitation of 17.4 mm to 42.0 mm.As far,it faces sand storms and dry-hot wind disasters.

Vegetation in the basin spreads sparse,with the growing season is from May to September.The total plant coverage is low and vegetation types are Phragmites australis,Salix rubra,Nitraria tangutorum,Achnatherum splendens,etc.Oasis in Tarim Basin is the main carrier for human life,with 42 counties in five states and four construction corps about 12×106peoples (Wang XY et al.,2017).

3.Data sources and methods

3.1.Data sources

Based on the data that the study had measured,the target literature was collected from CNKI and Science Citation Index.Colleting published papers about vegetation types and extreme groundwater level by using Groundwater depth,Western arid region,Tarim Basin,Groundwater-dependent ecosystem as keywords.Totally,1037 groups of groundwater depth data were collected which have 667 groups of measured data were distributed in 13 groundwater level observation wells (Fig.1; Supplementary Tables).

The vegetation data that have been used in this paper is from the MODND1F dataset,which is provided by the Geospatial Data Cloud site,Computer Network Information Center,Chinese Academy of Sciences (http://www.gscloud.cn).This dataset is synthesized from the MODND1D dataset with maximum values per five days.The spatial resolution of the dataset is 500 m.Collecting the normalization difference vegetation index (NDVI) data from 2000 to 2016,choosing the maximum data for the year to represent the vegetation coverage in the study area of a year.

3.2.Methods

The vegetation index is one of the indicators for vegetation analysis,which is combining the spectrum with data by using the sensitivity of certain bands of satellite to vegetation (Ma M and Veroustraete F,2005).The vegetation index includes simple vegetation index,ratio vegetation index,normalized vegetation index,conversion standardized vegetation index,etc (He L et al.,2013).In this paper,theNDVIwas used as the analysis index,the equation is as follows.

WhereNIRis TM-polarization near-infrared band value andRis the value of theTMvisible-infrared band.

TheNDVI’s range is generally −1 ≤NDVI≤ 1,and the positive value indicates vegetation cover increases with the increase of coverage; when the value is 0,it means that there are rocks or bare soil on the surface; the negative value indicates that the vegetation cover with clouds,water,snow,etc (Zhao YS,2003).

The downloadNDVIdata from 2000 to 2016 were analyzed by univariate linear regression,and the regression equation is as follows.

The calculation resultKcan be used to represent the maximum trend ofNDVIin one year.K> 0 means the increasing trend of vegetation coverage,andK< 0 shows that the regional vegetation coverage is decreasing.When theK≈0,it means that the change of vegetation coverage is stable.

The Pearson correlation coefficient is used to measure the correlation of vegetation phenology and meteorological data at monthly scale,and the equation is

Where thexiandyiare the climate and vegetation phenology data iniyear.The greater the absolute value ofRis,the greater the impact of climate indicators on plant growth is,and vice versa.

4.Results and discussion

4.1.Spatial and temporal distribution of vegetation coverage

In this paper,theNDVIdata from 2000 to 2014 were used to vegetation cover.the maximum value in a year were choosen to represent the vegetation cover of the year.TheNDVIdistribution in the study area is shown in Fig.2,and surface vegetation cover types are shown in Fig.3.

Fig.1.Groundwater monitoring wells distribution in Tarim and Qaidam basins,China.

In general,from 2000 to 2014,the vegetation coverage in Tarim Basin decreased initially and was followed by an increase.There was distinct vegetation degradation in the basin from 2000 to 2004 with the area proportion ofNDVIvalue > 0.5 decreased from 12 % to 6 %.It can be seen from Fig.2 that the most serious area of vegetation degradation is the desert area located in the center of the basin.TheNDVIvalue of the desert area decreased from 0.202 to 0.003.The result also showed that vegetation slightly affected in the mountain area around the basin,piedmont plains,and riparian.Whereafter,the regional averageNDVIvalue increased from 0.45 to 0.57,from 2004 to 2008.The vegetation in the basin began to recover gradually with area proportion ofNDVIvalue > 0.5 increased from 6 % to 12 %.The total vegetation coverage level in the Tarim Basin was similar to the 2000’s.However,the vegetation coverage was still low in the center of the basin,with an averageNDVIof 0.002.The desert area withNDVI< 0.05 increased from 1.2 % to 2.8 %.Meantime,vegetation in the western and southern mountainous areas began to degenerate,which caused the regional averageNDVIvalue to decrease from 0.23 to 0.11.The total desert area in the basin showed a decreasing trend,by the proportion of desert area withNDVI< 0.05 decreased from 2.32 % to 1.21%.From the whole spatial distribution change of vegetation index,the high-value area of vegetation index is the cultivated land,mostly in the piedmont alluvial plain area of the basin,which is the main area of human activities and lays the material foundation for oasis development.Since 2000,the desert area of the Tarim Basin had increased by 18 × 103km2,and the cultivated land area had increased by 16.3 × 103km2,with the proportion increasing from 64.43 % and 2.27 % to 66.41 % and 4.51 %.

Fig.2.NDVI distribution in Tarim Basin from 2000 to 2014.a-2000; b-2004; c-2008; d-2014.

Fig.3.Surface vegetation cover.

Qaidam Basin and Tarim Basin are both arid ecosystems,characterized by sparse vegetation and uneven distribution.Based on the above analysis of vegetation cover change in Tarim Basin for a long time scale (14 a),the study of Qaidam Basin mainly focuses on the changes of vegetation cover within the basin in a short time scale.Using the MODND1F data from 2014 to 2016,choosing the maximum value in a year to represent the vegetation cover of the year.The results ofNDVIchange are shown in Fig.4.

Fig.4. NDVI distribution in Qaidam Basin from 2014 to 2016.a-2014; b-2015; c-2016; d-the proportion of green space and desert area.

TotalNDVIdistribution in the Qaidam Basin shows that the vegetation coverage is low.The averageNDVIvalue of the most area in the basin is 0.05-0.15,the area proportion ofNDVIvalue < 0.05 is 3.10%,and the area proportion ofNDVIvalue > 0.5 is 11.28%.Area ofNDVI> 0.15 is mainly distributed in the northern,eastern,and southern mountainous areas and alluvial-pluvial fan areas.According to previous research findings,eastern and southern mountainous areas of the basin have high altitudes,mainly plateau meadows.Based on this,the growth of vegetation in the mountain area is good.Alluvial fan areas in the west and north are mostly the front of fluvial alluvial fan and lacustrine plain,where water and soil conditions are conducive to vegetation development and oasis formation.TheNDVIdata from 2014 to 2016 were analyzed by univariate linear regression analysis,and the analysis results were divided into five grades.The classification results are shown in Table 1.

Table 1.Grade table of regression analysis results

Vegetation around East Taigener Salt-lake degraded conspicuously,and theNDVIdecreased from 0.596 (2014) to 0.005 (2016),K= −0.295.The vegetation changes in the northern,eastern,and southern mountainous areas and alluvial-pluvial fan areas of the basin are relatively stable,with aKvalue of 0.02.Overall,the area proportion ofKvalue from 0.2 to 0.5 accounted for 3.1 %,theKvalue is 0.2-0.5 area accounted for 0.8 %,and the remainingKvalues are in the range of −0.2 - 0.2.The obvious area of vegetation degradation is concentrated in the northwest basin with a lowNDVI.

4.2.Analysis of vegetation change during degradation period from 2000 to 2004

Previous research shows that vegetation is sensitive to climate change,especially to precipitation,temperature and solar radiation (Wei ZF et al,2014; Zeng B and Yang TB,2009).Groundwater is the dominant water source for vegetation to maintain life activities in the arid area.Therefore,the change of groundwater depth has an important impact on vegetation type and ecological structure (Robinson DA et al.,2008).The analysis results of NDVI change in Tarim Basin from 2000 to 2014 proved that,there was obvious vegetation degradation in Tarim Basin from 2000 to 2004.The areas with the most obvious degradation concentrated in the center of desert area of the basin.On the contrary,the degradation degree of vegetation in the surrounding mountains and rivers is slight.To explore the relationship between vegetation degradation and precipitation,groundwater in the study area.The monitoring wells of Kongque River with abundant data were selected as the research area.Then,the average annual NDVI,annual precipitation and depth to groundwater table were analyzed by linear regression analysis.The analysis results are shown in Fig.5.

The results of correlation analysis showed that there was a negative correlation betweenNDVIand the changes of precipitation and groundwater depth during vegetation degradation period from 2000 to 2004.TheNDVIin Kongque River catchment decreases with the increase of groundwater depth and precipitation.Farthermore,the change of depth to groundwater table has a strong correlation withNDVIindex,and the absolute value of correlation index is 0.505.Precipitation has little effect on surface vegetation,and the absolute value of correlation index is 0.042.Statistical Analysis of annual precipitation data in Kongque River catchment from 2000 to 2004 shows that,the average precipitation in four years is 98.92 mm and maximum precipitation is 141.5 mm (2003).Under these conditions,precipitation cannot maintain the growth of vegetation.Therefore,groundwater becomes the main source of water for the growth of surface vegetation in the region.

Fig.6 shows that from 2000 to 2004,the change of groundwater andNDVIin the malachite river basin was opposite.when the depth to groundwater table increased,theNDVIvalue decreased.It can be seen from Fig.4 that the high value ofNDVIin the Kongque River catchment is mainly concentrated in the water level depth of 2 m to 5 m.The minimum value of regionalNDVIcorresponds to the depth to groundwater table of 7.5 m.Generally,the fluctuation of groundwater level in the Kongque River catchment is the main reason for the change of vegetation coverage during vegetation degradation from 2000 to 2004.

4.3.Analysis on vegetation change in recovery period from 2005 to 2014

Vegetation in the Tarim Basin recovered gradually from 2005 to 2014 and returned to the surface vegetation coverage level as 2000 in 2008.Analysis of influencing factors of surface vegetation change from 2000 to 2004 shows that vegetation growth is mainly restricted by groundwater level during vegetation degradation in basin.Due to small amount of annual precipitation in the basin,precipitation has slight effect on vegetation change in the region.To further analyze the influencing factors of surface vegetation change in vegetation recovery period,four meteorological stations,Kashi,Kurgan,Aksu and Bayinbuluke,were selected to cover mountainous and plain areas in the basin.Statistical analysis of daily meteorological data recorded by meteorological stations from 2005 to 2014.The result shows that,average precipitation in Tarim Basin from 2005 to 2014 was 49.42 mm.The maximum annual average precipitation was 75.1 mm,showing an overall increase trend year by year.Under this condition,the linear regression analysis of regionalNDVIwith depth to groundwater table and precipitation is conducted (Fig.7).

As can be seen from Fig.7,precipitation in the study area increased yearly,which enhanced the correlation between precipitation and regionalNDVIvalue,with the correlation coefficient as 0.328.Meantime,the correlation between groundwater and regionalNDVIwas weakened,and the correlation coefficient was −0.248.In general,under the trendof increasing precipitation,the dependence of surface vegetation growth on groundwater is significantly weakened.In order to further explore the relationship between vegetation growth and depth to groundwater table,theNDVIdata in the central basin from 2004 to 2008 with the largest increase in annual precipitation were selected for correlation analysis with depth to groundwater table.

Fig.5.Spase correlation analysis of annual precipitation,groundwater and NDVI in Kongque River catchment.a-the realationship of precipitation and NDVI; b-the realationship of goundwater depth and NDVI.

The high-value point ofNDVI(NDVI> 0.5) in the middle of the basin is concentrated in the range of 0 to 4 m of groundwater depth,and the overallNDVIvalue decreases with the increase of groundwater depth.The existing research results on species diversity in the west of Tarim Basin show that the dominant families of the main flora in the west of Tarim Basin are Chenopodiaceae,Compositae,and Leguminosae,and the representative plants arePopulus euphratica,Salix psammophila,Phragmites australis,Camelhornia,Potentilla,etc (Han L,2014).Due to developed root system,arbor and shrub are less affected by groundwater level fluctuation in shallow depth.On the contrary,herbaceous plant roots are relatively short and sensitive to groundwater level changes.When the depth to groundwater table in the region increases,the growth of herbaceous plants on the surface is gradually affected and eventually withered,which resulting in the decrease of regionalNDVI.As shown Fig.8,when the depth to groundwater table exceeds 2 m,the regionalNDVIdoes not decrease significantly.The main reason for this phenomenon is that the dependence of vegetation on groundwater is weakened because of the increase of regional precipitation.In other words,the increase of precipitation alleviates the water shortage of vegetation growth in the region to a certain extent.

Fig.6.The relationship of groundwater depth and NDVI in the Tarim Basin,China.a-relationship of groundwater depth and NDVI;b-the relationship of groundwater depth and NDVI in the central of the basin; c-the relationship of groundwater depth and NDVI in the east of the basin.

Fig.7.Correlation analysis of NDVI with depth to groundwater table and precipitation in Tarim Basin from 2005 to 2014.a-the relationship of Precipitation and NDVI; b-the relationship of depth to groundwater table and NDVI.

Fig.8.Relationship between depth to groundwater table and NDVI in central tarim basin from 2004 to 2008.

5.Conclusions

The total vegetation in the study area decreased at first and followed by an increase from 2000 to 2014.Severe vegetation degradation was observed in the part of the study area.In Tarim Basin from 2000 to 2004,there was obvious vegetation degradation with the area proportion ofNDVIvalue > 0.5 decreased from 12 % in 2000 to 6 % in 2004.The most server vegetation degradation occurred in the desert area,with theNDVIvalue decreased from 0.202 to 0.003.Subsequently,vegetation was gradually restored from 2004 to 2014,and the regional averageNDVIvalue was restored to the same level as in 2000 in 2008.The total coverage of Qaidam Basin is low,withNDVI< 0.05 area accounting for 3.101 %.Vegetation around East Taigener Salt-lake degraded significantly,NDVIdecreased from 0.596 (2014) to 0.005(2016),with theKvalue of −0.295.

(ii) The fluctuation of groundwater level is the main reason for the change of vegetation coverage during the vegetation degradation in the basin.However,the average annual precipitation in the study area is less,which is not enough to have a significant impact on vegetation growth.The groundwater depth in the study area have strong correlation withNDVI,and the absolute values of correlation index is 0.505.The change trend ofNDVIindex is opposite to that of depth to groundwater table.Namely,that the depth to groundwater table increases and theNDVIindex decreases.

(iii) The annual average precipitation showed an increase trend during the vegetation restoration in the basin.Under these conditions,the dependence of vegetation on groundwater is obviously weakened with the correlation index−0.248.The correlation between precipitation and regionalNDVIwas enhanced,and the correlation coefficient was 0.328.The increase of precipitation alleviates the water shortage for vegetation growth in the region.

CRediT authorship contribution statement

Jin-ting Huang conceived of the presented idea.Ge Song and Bo-han Ning verified the analytical methods.Jin-ting Huang encouraged Ge Song and Bo-han Ning to investigate groundwater levels and supervised the findings of this work.All authors discussed the results and contributed to the final manuscript.

Declaration of competing interest

The authors declare no conflicts of interest.

Acknowledgment

This research was supported by the National Natural Science Foundation of China (41672250,42177076),the Natural Science Foundation of Shaanxi Province (2019JLZ-03),and the Key R & D Projects of Shaanxi Province(2021ZDLSF05-09).

Supplementary data

Supplementary dataset (Table S1,Table S2 and Table S3)to this article can be found online at doi: 10.31035/cg2021062.