Responses of distribution pattern of desert riparian forests to hydrologic process in Ejina oasis

Science in China Ser. D Earth Sciences 2004 Vol.47 Supp.I 21- -3121Responses of distribution pattern of desert riparian foreststo hydrologic process in Ejina oasisZHAO Wenzhi 1, CHANG Xueli 2 & HE Zhibin'1. Linze Inland River Basin Comprehensive Research Station, Cold and Arid Regions Environmental and Engineering Research Institute,Chinese Academy of Sciences, Lanzhou 730000, China;2. Yantai Normal University, Yantai 264005, ChinaCorrespondence should be addressed to Zhao Wenzhi (email: zhaowzh@ ns.zb.ac.cn)Received September 5, 2003Abstract By using the theories and methods of landscape ecology and the technology of GISand RS , a study has been carried out on the responses of distribution pattern of desert riparianforest to hydrologic process on the basis of the hydrologic data from 1990 to 2000 and the TMimage of 2001 year. The results showed that: (1) there appears an even distribution pattern forthe relative forest area in oasis, however, the degenerated forest diaplays an increasing ten-dency from west to east; (2) the desert riparian forest in Ejina is in completely degeneratedprocess at the patch scale; (3) the number of patch is influenced not only by hydrologic process,but also by agricultural activity such as cultivation. The severe deterioration of the degradedvegetation in whole oasis initiates from lower reaches, and gradually impels to upstream; thefragmentation of landscape in the terminal site is more obvious, which is influenced by rivershape and decreasing flux of water. It is found that the influence of surface hydrologic process tothe ground hydrologic process of desert riparian forest in Ejina oasis is ltte for the recent tenyears. The relative area of the degenerated forest increased with increasing ground water depthin the direction of parallel to river channel. On the contrary, in the direction perpendicular to riverchannel, there is a decreasing tendency for the average patch area of the forest and the degen-erated forest with increasing ground water depth.Keywords: Ejina oasis, desert riparian forests, vegetation pattern, hydrologic process.DOI: 10.1360/04zd0003Riparian corridors are critical habitat for deserttion in vegetation pattemn, and vegetation pattern alsoflora and fauna, providing oases of species diversityresponds to the change of hydrologic process'and productivity in otherwise dry environments"'. InThe Ejina desert riparian forest is an importantarid regions, plants in the desert riparian forest, espe-cially non-zonal mesophytes representative of desertcomponent of the Heihe River basin, the second larg-est inland river basin in China. It not only has signifi-riparian species, depend on underground water ansurface water, besides rainfall'2- . Therefore, surfacecant hiolggical function to the biodiversity conserva-and underground hydrologic processes control thetion中国煤化工important significancepattern and composition of the desert riparian forests.to mHCNMHGftheEjinaregionandThe change in hydrologic process can lead to altera- the Heihe River basin'". In the recent decades, theCopyright by Science in China Press 200422Science in China Ser. D Earth Sciencescoming water amount has decreased year by year in1 Study area, materials and methodsthe Ejina area in the lower reaches with the increasing1.1 Study areaneed for water in midstream and upstream8. The lim-ited water resource reaching Ejina desert riparian for -The study area is located in the Ejina oasis of theest is controlled by both human and natural factors. Inlower reaches of the Heihe River, extending from thethe direction parallel to river channel, the constructedcenter of Ejina banner to the northeast of Dalahubu8 water gates as water channels had been used to con-town (41930'- -42*30N, 99*45'一101930E). This areatrol the distribution of water resource all the times. Inbelongs to the alluvial plain of the Ejina River, andthe direction perpendicular to river channel, waterdeclines gradually from southwest to northeast, withflow is natural rather than controlled by human. Influ-an average elevation of about 1000 m. The climate isenced by natural and human factors, the primarycharacterized by drought and less rain with a coldvegetation patterm of the Ejina desert riparian forestwinter and hot summer, and high wind and more sand.dominated by Euphrates Poplar has considerablyIt has obvious characteristics of four seasons, and achanged. The standing forest showed severe deteriora-great temperature difference between day and night.tion, and was characterized by severe unbalance ofThe average annual temperature arranged from 6 tostanding forest age structure and decrease of Euphrates8.5C, averaging - 09一- 14C in January, with thePoplar area. The old individuals occupy a large pro-minimum of - 36.4C, 22一26.4C in July with theportion and the young and mid-aged individuals onlymaximum of 41.7C. The annual average frost-freebuild up a small part of the standing forest.period is 130一165 days. Mean annual precipitationis 40 mm, and annual evaporation is 4200 mm. MeanAt present, numerous studies have been carriedannual wind velocity ranges from 2.9 to5 m.s，andout concerming the responses of desert riparian forestsprevailing wind direction is in northwest. Gales witho hydrologic process, which mostly focused on thethe velocity≥>17 m. s-' occur in about 70 days. Theresponse of desert riparian forests to groundwaterzonal soils are mainly gray desert soil and gray-brownlevel4.6.9 121 and the responses of desert riparian for-desert soil, with a small part of saline sodic soil andests to surface hydrologic process')n variousmarshy soil in lake basin and bottomland. Dominantscales such as individuals, populations and communi-plant species are Ejina desert riparian forest Populusties. On landscape scale, van Coller et al.!4 combinedeuphratica, some shrubs including Tamarix chinensis,gradients methods with patch hierarchy approaches toSophora alopecuroides, Elaeagnus angustifolia,determine the relationships between riparian vegeta-Haloxylon ammodendron and some herbaceous in-tion and environmental factors. However, no study hascluding Lycium ruthenicum, Phragmites communis,been carried out in this aspect in China.Poacynum hendersonii, Karelinia caspica, AlhagiIn this paper, we selected the gradient of surfacesparsifolia,Aneurolepidium dasystachys, Ach-hydrologic process controlled by human and naturalnatherum splendens, Suaeda SPP .surface hydrologic process in the Ejina desert riparianSurface water originated from the Heihe River,forest ecosystem to investigate the distribution patternand is called Ejina River after flowing into Ejina oasis.of oasis vegetation and changes of surface and under-The flow length of the Ejina River is about 250 km.ground hydrologic process and to analyze the responseAnnual runoff of the Ejina River was 1.2 to 1.3 billionof oasis vegetation patterm to hydrologic process usingm' in the 1950s, and flowed into east and west JuyanRS and GIS technology. The aim is to determine theSea. From the 1960s to 1970s, the runoff was at leastinteraction between desert riparian forest vegetation中国煤化工the runoff decreasedand hydrologic process from the viewpoint of eco-to 0.MYHCNMHG.0.2 billion m3 in lowhydrology.Responses of distribution pattern of desert riparian forests to hydrologic process in Ejina oasis23The data of ground water depth from 13 moni-south of east Juyan sea, covering an area of 10100 hm',toring points during 12 years in Ejina oasis ilustrateand with 11 sampling plots established 4000 m apartthat, the annual ground water depth ranges from 2.18from each other. It represents the vegetation distribu-to 3.11 m, with the least water depth of 2.18 m in Apriltion pattern from top to low parts of the oasis, and isand the greatest depth of 3.11 m in December, a dif-influenced by natural factors. With respect to the ac-ference of 0.93 m. There is a significant decreasingcuracy of the image interpretation and clarity of patchtrend for regional ground water depth compared to theboundary in reality, and relative stability of forest vi-average level of ground water depth in rich waterability in short time, we selected some variables re-years, and this reduction reaches about 1m.lated to forest as the analysis index for vegetation pat-tern analysis. These variables include relative wood-1.2 Materials and methodsland area (ratio of woodland area to study area, WA),(1) Image interpretation. The information isrelative deteriorated woodland area (ratio of deterio-from Landsat TM image of July 2001 in this study.rated woodland area to study area, DWA), averageErdas 8.4 was performed to use geometry calibration,woodland patch area (AWPA), average deterioratedand Arcview 3.2 and Arcinfo 7.0 were used to imagewoodland patch area (ADWPA), number of woodlandinterpretation. The land use and cover types was re-patches (WPN) and number of deteriorated woodlandgarded as the standard of interpretation, and was di-patches (DWPN).vided into 11 types (table 1, fig. 1).(3) Surface and underground hydrologic proc-(2) Pattern analysis. The transect method wasesses. Surface hydrologic process was analysedapplied to vegetation pattern analysis. Two transectsbased on the flow data of Zhengyixia from 1990 towere marked, one parallel to river channel (east- west2000. The analysis on underground hydrologic processdirection), and another perpendicular to river channelwas made based on the data of groundwater burial(south-north direction). The parallel transect was fromdepth during 1990 to 2000 at the three fixed monitor-the west of Dalahubu town to the east of the watering points, i.e, the herdsman drinking well in the vi-gates, covering an area of 3800 hm', and with 11 sam-cinity of No. 7 water gates (42.020N, 101.257° E), .pling plots established at the interval of 2000 m. ThisSuoguodiaoer artesian well (42.337° N, 101.248° E)transect can ilustrate the influence of human controlon vegetation distribution patterm. The perpendicularand no. 2 mountain protecting town river (41.928* N,one was from 3 km north of the point where the101.035° E). The data from the three monitoring pointsHeihe River branches towards east and west, to thecan be considered to represent most parts of the EinaTable 1 Classification of land use image interpretation in Ejina deltaType .FarmlandMainly including irigated field, salted irigated field, sandy irigated fieldPlantationMainly including orchard, nursery, arificial protecting forest, young woodlandRiparian forestMainly including Populus euphratica forest, Elaeagnus angustifolia forest, Tamnarix chinensis forest, andtheir various combinationsDesert forestMainly including Haloxylon ammodendronforest,Tamarix chinensis forest, and dfferet degree ofdegenerated forestsDegenerated forestMainly including degenerated Populus euphratica forest, Haloxylon ammodendron forest, Tamarix chinen-sis forest, etc.Mainly including Phragmites communis, Achnatherum splendens. Sophora alopecuroides, artificial grass-Riverine grasslandland, and s0 onDesert grasslandMainly including Lycium ruthenicum, Nitraria roborowskii, Alhagi sparsifolia, Karelinia caspica, Reau-muria soongorica, Ephedra przewalskii, and中国煤化工.Ip and bttomland, etc.Dried lake basin and river bed Mainly including dried lake basin, rservoir, seResidential areaMainly including town, village, mine, etc.iYHCNMHGGobiMainly including Gobi, mobile sand dune, bare land, saline-alk ali land and upland, etc.24Science in China Ser. D Earth Sciences101.00°101 .25°101.5°"42.25°| 42.00°Observation points of ground water懈Farmlandi:品Man-made forest1 Riparian forestm Desert grasslandDesert woodland2M Floodplain meadow8 Dry riverbed41.75°Ejina County口Desert州Degenerated woodland一Channel020 30 kmFig. 1. Land use status and location of ground water depth monitoring points in Ejina delta.desert riparian forests and the present study area. TheHeihe River, with an area of about 1300 km'. At pre-isobathiclines were generated by the 'Surface' functionsent, the flux condition of 8 branches is regulated byof Arcview 3.2 software. The groundwater depth ofthe upriver hydro-junction. Therefore, the vegetationeach plot was replaced by the average value of iso-distribution from No. 1 Bridge to No.8 Bridge can bebathiclines under the plot, for example, if the area of aregarded as the consequence in response of humanplot overlapped the isobathiclines of 3.2 m, 3.3 m, 3.7regulation to hydrologic process. The information ofm and 3.8 m, then the average groundwater depth ofdistribution pattern can be obtained from the transectthe plot should be 3.5 m.parallel to the river.2 ResultsThere is an increasing tendency gradually fromwest to east for WA and DWA (fig. 2). In the central2.1 Responses of vegetation pattern to human-regu-part of the region (No. 1 Bridge to No. 8 Bridge cor-lated hydrologic processrespond to the third plot to the eleventh plot), WAThe Heihe River branches into the east river andchanges from 47.5% to 91.7%, with the highest valuethe west river after it flows into the Ejina oasis fromof 91.7% at 10 km from the initiated point (fig. 2(a),Zhengyixia. The east river is subdivided into 8the perpendicular arrow. between No.1 Bridge and No.branches, showing a braided water system. The drain-2 Brid中国煤化工e of 88.9% at 19 kmage area of east river formed the biggest oasis, the (fig. 2|YHC N M H Gs euphratica conser-Ejina desert riparian forest, in the lower reaches thevation district, as pointed by horizontal arrow). ThereResponses of distribution pattern of desert riparian forests to hydrologic process in Ejina oasis25is only a peak value for DWA (fig. 2(b), as pointed byplot. There is an increasing trend from west to east,perpendicular arrow) at 19 km (No. 7 Bridge Populusand the increasing gets faster in both woodland andeuphratica conservation district), which is coincidentdegenerated woodland 12 km away from No.1 Bridge.with the second peak value of WA.ADWPA is greater than AWPA, which suggests thatBasically, there is a similar trend in the change ofthe degenerating trend of desert riparian forest appearsdistribution pattern for AWPA and ADWPA (fig. 2(c),as a whole, more severe degeneration occur in the east(). The difference is that ADWPA is greater thanof woodland, and the last plot (east of No. 8 Bridge)AWPA, except the same area of 720.1 hm2 in the lasthas degenerated completely.10.0100.0py= 20.668.n(x)+14.925R2= 0.48080.0-80.060.0-60.040.0y = 20.668Ln(x)+14.925R2=0.48020.0-(b0.0L0.01020 2515Distance/km900750-y= 2.506x2 -42.241x +174.24750y= 2.65.2 -4.92x +191.39.R2=0.772R2= 0.827600450ξ 450300150c)(d)005115 20 25051012050.0 r8(y= -0.472x2 +11.027x-11.523y= -0.288x2 +6.975x -9.926R=0.708R=78960}30.0 |爹40-(。(fo.0LDistance /km中国煤化工Fig.2. Ditribution pattern of desert riparian forest infuenced by human-regulatMHC N M H Gwoodland area, (b) relativedegenerated woodland area, (C) average woodland patch area, (d) average degeneratea woodland patcn area, (e) numder of woodland patches, (I)number of degenerated woodland patches.26Science in China Ser. D Earth SciencesWith respect to the distribution pattern of wood-2.2 Responses of vegetation to natural hydrologicland patch number, the biggest value of patch numbersprocessin both woodland and degenerated woodland occur atThe flowing direction of the Heihe River is12 km away from No.1 Bridge and average 84 and 45,south-north in Ejina desert riparian forest, so the plotsrespectively. The area with the biggest patch number isthat perpendicular to river can be taken to reflect re-located in the middle part of the oasis (in the vicinitysponse of vegetation to the river natural hydrologicof No.3 Bridge and No.4 Bridge).process from upstream to lower reaches.80.0r80.070.0y= 0.951x+25.661y= 0.973x+12.644R2= 0.748R2=0.65660.050.0-50.0s 40.040.0 t30.020.0 t●20.0上0(b)10.01030 40 5002030405Distance /kmDistance/km360y= 0.478x2 -17.685x +221.52310 y= 0.272x2 -9.151x +143.7510R2=0.74R2= 0.821410210-310160210 t60(C110(d203040 5010 203050Distancekm80p4570-y= -0.083x2 +3.918x +16.20340R =0.78060-3550-25 ty= -0.044x2 +2.299x +7.387e)R2= 0.62324(203(Fig. 3. Distribution pattern of desert riparian forest infuenced by the natural h中国煤化工Jland area, (0) rlative deger-ated woodland area, (C) average wodland patch area, (d) average degenerated wdsodland patches, () number of:YHCNMHG^degenerated woodland patches.Responses of ditribution pattern of desert riparian forests to hydrologic process in Ejina oasis27There is a gradually increasing trend from southdata analysis of the direction parallel to river haveto north for both WA and DWA along the transect (fig.several characteristics. First, with respect to the dis-3). WA ranges from 19.7% to 74.7%, with the highesttribution pattern of woodland and degenerated relativevalue at about 46 km away from the initiate point inwoodland area, there occurs an even distribution pat-the last plot, ilustrating that WA in Ejina desert ripar-tern for the relative forest area in oasis (fig. 2(a)).ian forest is gradually increasing from upstream toHowever, there is an increasing trend from west to eastlower reaches in oasis. DWA ranges from 13.3% tofor the distribution patter of degenerated woodland70.3%, with the lowest value of 13.3% in the forth plot(fig. 2(b)). Second, the distribution pattern of average(about 15 km).patch area for the woodland is similar to that for theThere is a similar trend in the change of distribu-degenerated woodland, suggesting that at present thetion pattern for AWPA and ADWPA, exhibiting aEjina desert riparian forest is in completely degener-fluctuated increase (fig. 3(c), (d)). The lowest valueated process at patch scale. Third, in view of the dis-and the highest value appear at the forth plot and thetribution pattern of patch number, the middle region ofthe Ejina desert riparian forest (at 10 to 12 km) is in-eleventh plot, respectively. However, the range offluenced not only by hydrologic process, but also bychange differs between AWPA and ADWPA: AWPAagricultural activities, showing a high patch fragmen-from 54.8 to 316.8 hm, ADWPA from 48.7 to 516.8tation in this region(fig. 2(e), ()).hm'. In addition, in the range from 0 to 37 km (fromthe first plot to the ninth plot, within the imaginaryInfluenced by natural hydrologic process, theframe of fig.3 (C), (d)), the variation range is litte forcharacteristics of the patch areas of woodland and deboth AWPA and ADWPA. However, in the range fromgenerated woodland show that the effect of natural37 to 46km, considerable alteration is observed, andhydrologic process is more significant in the lowerADWPA reach 516.8 hm2 in the north of oasis.reaches than in upstream, and that the number of de-generated woodland patches increased significantly atThe highest value of woodland patch numberthe end of river oasis. This is related to the shape ofoccurs at 20 km away from the initiate point, reachingthe lower reaches of the Heihe River because the in-78 in number; the highest value of degenerated wood-creasing branches in the lower reaches led to segmen-land patch number appears at 37 km away from thetation of woodland, and, in turn, increase in the num-initiate point, 43 in number. In the adjacent region 3ber of patches. Subsequently, forest is degeneratedalmost equal peak values continuously appear withinwith the degeneration of hydrologic condition. At thethe range of 29- -37 km (fig. 3()).end of the river, the branches sink to one and flow intoeast Juyan Sea, resulting in a decrease in the number3 Discussionof patches (fig. 3(e)). Correspondingly, the patchHydrologic process is the most principal naturalnumber of degenerated woodland forest decreases.factor to affect the patter of desert riparian forests.However, there are the largest areas for woodland andSurface and underground hydrologic processes deter-degenerated woodland within this region (fig. 3(a),minate spatial pattern of vegetation in oasis(11.12](b)).Change in vegetation reflects the alteration of hydro-The influence of natural hydrologic process onlogic process, especially for riparian forest with highstability compared to other vegetation type such asdesert riparian forest reflected in the direction perpen-grassland. Therefore, change in the riparian forest isdicular to river is expressed as the following aspects.the consequence of hydrologic process on a relativelyFirst, the areas of both woodland and degeneratedlong time scale.wood中国煤化工mbers are increasingInfluences of human- regulated hydrologic proc-fromYH: that the degenerationCN M H Gin the lower reachesess on Ejina desert riparian forest resulted from theand developed to upstream (fig. 3(a)- ()), This deg-28Science in China Ser. D Earth Sciencesradation process is accompanied with the decreasingrather than the supplied groundwater at present. Theflux of the Heihe River. Second, the influence of natu-supplement of surface water to groundwater is far lessral hydrologic process on woodland landscape frag-than water consumption of oasis. Therefore, the fluxmentation of the Ejina desert riparian forest occursfluctuation from Zhengyixia can not lead to the fluc-mainly in the terminal part of the Heihe River. Thistuation of underground water level in oasis.displays as a consequence of river shape and decreas-ing flux.and underground hydrologic processThe responses of desert riparian forest to surfacehydrologic process regulated by both human andZhengyixia flux/m' .s'-0.178-0.085natural factors have common characteristics, which:an be indicated by forest area and patch numbers.Unlike the surface hydrologic process, the under-Woodland and degenerated woodland areas, averageground hydrologic process has close relationship withwoodland and degenerated woodland patch areas,the vegetation of asis, and there is significantpatch numbers increased with increase in distance.relationship between desert riparian forest and under-Generally, surface hydrologic process has high vari-ground hydrologic process on long time scale. In theability, and the similar regularity of vegetation patternpresent study, the data of ground water depth for 11in east- west direction and south- north directionyears from the three monitoring points (A, B, and C)means that pattern of desert riparian forest is con-can be taken to reveal the ground water depth in thetrolled by both surface hydrologic process and under-core region of the Ejina desert riparian forest. Withground hydrologic process. The influence of surfaceGIS technology, corresponding analysis on distribu-hydrologic process on desert riparian forest may betion patterm of ground water depth in the core regionrelated to reproduction and regeneration of Populuscontrolled by three points (fig. 4) and distribution pat-euphratica, however, underground hydrologic processtern of each variable of oasis woodland can reveal theis related to their growhl5.16.influence of underground hydrologic process on pat-Since East Juyan Sea dried in 1992, very lttletern of oasis woodland. The analysis results indicatewater of the Heihe River flew through Ejina desertthat (table 3), in the direction perpendicular to river,riparian forest, therefore, there is litle direct influencethere is a greater influence of underground bydrologicof surface hydrologic process on vegetation of Ejinaprocess on relative woodland area, relative degener-desert riparian forest, and the effect of hydrologicated woodland area, average woodland patch area andprocess on vegetation resulted mainly from ground-average degenerated woodland patch area, and awater supplied through surface water. Therefore, thesmaller influence on patch number of woodland andrelationships between surface and underground hy-degenerated woodland. It is noted that the influence ofdrologic process and between groundwater processunderground hydrologic process on the relative de-and vegetation patterm should be emphasized.generated woodland area is more significant, with theCorrelation analysis between the flux from Zhen-correlation coefficient of 0.833. This suggests that inthe direction perpendicular to river, the degeneratedgyixia (representative of surface hydrologic process)woodland is significantly increasing in patch scaleand the data of ground water depth from threewith increase of the underground water depth.monitoring points during 1990 to 2000 (table 2) indi-cates that there is lttle influence of surface hydrologicIn the direction parallel to river, underground hy-process on underground hydrologic process of thedrologic process exerts significant influence on AWPA,Ejina desert riparian forest. The maintenance of theADV中国煤化工atively correlated withEjina desert riparian forest for the recent ten years de-MHCNMHGcorrelated coefficientspends on consumption of the stored groundwaterof -0.041 anu-- U.ouL, lespectively. This illustratesResponses of distribution pattern of desert riparian forests to hydrologic process in Ejina oasis29Table 3 Correlation coefficients between ground water depth and distribution patterm of woodlandWADWAAWPAADWPAWPNDWPNPerpendicular riverway0.5070.8338)0.4310.4530.0680.222Parallel riverway-0.386-0.581-0.841*-0.802)0.794)0.590a) Signifcance at 0.01 level●3.60，4.2004.05496_| 3业●Observation points ofground water] Sampling areaIsolines of groundwater depthFig. 4. Superposition of vegetation pattern and isolines of ground water depth in the study area.that AWPA and ADWPA decrease significantly withthe woodland patch fragmentation caused by under-the increase in underground water depth (fig. 5 (b)-ground hydrologic process.(). The smaller the average patch area, the higher the4 Conclusionsdegree of landscape fragmentation (the more thenumber of patches) in a given studied area"", which(1) River hydrologic process in arid zone is in-can be confirmed by the positive correlation betweenfluen中国煤化工latural factors. In theWPN and DWPN and underground hydrologic processprese|YHCNMHGalleltoriver(ineast-as listed in table 3. This alteration can be ascribed towest direction) and perpendicular to river (in south-30Science in China Ser. D Earth Sciences100.0 ry= 85.683x -282.93300.0R2= 0.693●80.0y--1.76.92*+812.56 .250.0R2= 0.70760.0200.0 t40.0，150.0 t20.0100.0 t:●、●(a(b)3.403.804.00 4.203.203.403.603.804.00 4.203.40° 3.60Ground water depth/m60.0p450.0y= 3.404x-81.878y= -274.28x+1218.550.0 tR=0.630R2= 0.643350.0 t250.0 t屋40.0受30.0.,●50.020.0 L 23.00 3.20 3.403.60 3.80 4.00 4.203.003.203.403.603.804.00 4.20Fig. 5. Relationship between vegetation pattern and ground water depth. (a) Relative degenerated woodland area and ground water depth in thetransect perpendicular to river channel; (b) average woodland patch area and ground water depth in the transect parallel to river channel; (C) averagedegenerated woodland patch area and ground water depth parallel to river channel; (d) number of woodland patches and ground water depth in thetransect parallel to river channel.north direction) reflect the response of vegetation towith increase in ground water depth in the directionsurface hydrologic process. On the east- west transect,parallel to river channel. Contrarily, there is a de-there appears an even distribution pattern for the rela-creasing tendency for the average patch area of thetive forest area in oasis, however, the degenerated for-forest and the degenerated forest with increase inest shows an increasing tendency from west to east.ground water depth in the direction perpendicular toThe desert riparian forest in the Ejina is in completelyriver channel.degenerated process on patch scale. The patch numberAcknowledgements This work was supported by Grand Project ofis influenced by both hydrologic process and agricul-Knowledge Innovation Program of Chinese Academy of Sciencestural activities. On the south- north transect, the de-(Grand Nos. KZCX3-SW-329 and KZCX1-09) and National NaturalScience Foundation of China (Grand No. 40235053).generated trend of oasis vegetation started from lowerreaches, and gradully developed to the upstream. In-Referencesfluenced by river shape and flux decreasing, landscape1. Poff, N., Allan, J, Bain, M. et al. The natural flow regime, Bio-fragmentation is more obvious at the termination ofscience, 1997, 47: 769- -784.the river.Mensforth, L. J. Thorburn, P. J. Tyerman, s. D. et al, Sources of(2) The influence of surface hydrologic processwater used by riparian Eucalvotus camaldulensis overlying highly中国煤化工94, 102--28.on the ground hydrologic process of desert riparianforest in the Ejina oasis is lttle for the recent ten years.YHCNMH_imith, S. D. Water uptake in ahe southwestern United States: aThe relative area of the degenerated forest increasesstable study, Ecological Application, 1992, 2: 450- 459.Responses of distribution pattern of desert riparian forests to hydrologic process in Ejina oasis314. Chen Yaning, Chen Yapeng, Li Weihu et al, Response of the ac~11. Chen Yaning, Li Weihong, Xu Hailiang et al., The influence ofcumulation of proline in the bodies of Populus euphratica to theunderground water level to vegetation in Tarim River lowerchange of groundwater level at the lower reaches of Tarim River,reaches, Acta Geographica Sinica (in Chinese), 2003, 58 (4): 542Chinese Science Bulletin, 2003, 48 (18): 1995- 1999.. Jansson, R., Nillson, C, Dynesuis, M. et al, Effects of river regu-12. Zhong Huaping. Liu Heng. Wang Yi et al., The relationship oflation on river-margin vegetation: a comparison of eight borealEjina desert riparian forest and water resource in Heihe lowerrivers, Ecological Application,, 200. 10(1): 203- 224.reaches, Advances in Water Science (in Chinese), 2002, 13 (2):6. Horton, J. L., Clark, J. L., Water table decline alters growth and223- 228.survival of Salix gooddingu and Tamarix chinesis seedlings, For-13. Zamora-Aroya, F, Nagler, P. L. Briggs, M. et al, Regenerationest Ecology and Management, 2000, 140: 243- 251.of native trees in response to flood releases from the United States7. Liu Zhongliu, Zhu Zongyuan, Hao Dunyuan, Thinking about 0a-into delta of Colorado River, Maxico, Journal of Arid Environ-sis ecosystem damage and strage of ecology conservation inments, 200, 49(1):49- -64.Heibe River lower reaches, Journal of Arid Land Resources andvan Coller, A. L, Rogers, K. H, Heritage, G L, Riparian vegeta-Environment (in Chinesc), 2001, 15(3): 1- -8.tion- environment relationships: complimentarity of gradients8. Wang Genxu, Cheng Guodong, The changes of hydrography andversus patch hierarchy approaches, Joumal of Vegetation Science,ecological environment of Heihe River basin in recent 50 years,2000, 11:337- -350.Joumal of Desert Research (in Chinese), 1998, 18(3): 233- -238.15. Forest Editing Committee of Inner Mongolia Autonomous Region,9. Horton, J. L, Kolb, T. E.. Hart, s. C, Responses of riparian treesForest of Inner Mongolia Autonomous Region (in Chinese), Bei-to interannual variation in ground water depth in a semi-arid riverjing: China Forestry Publishing House, 1989.basin, Plant Cell and Environment, 2001, 24: 293- -304.16. Huang Peiyou, Non-imigation Vegetation and Restoration in Arid10. Stromberg, J. C, Tllr, R.. Richer, B., Effects of ground waterZone (in Chinese), Beijing: Science Press, 2002.decline on riparian vegetation of semi-arid zones, The San Pedro,7. Forman, R. T. T, Godron, M., Patches and structural componentArizona, Ecological Application, 1996, 6: 113- -131.for a landscape ecology, Bioscience, 1981, 31: 733- -740.中国煤化工MYHCNMHG