简介：Themethaneconcentrationprofilefrom-1.5mdepthinsoilto32mheightinairwasmeasuredinalpinesteppelo-catedinthepermafrostarea.Methaneconcentrationsshowedwidelyvariationsbothinairandinsoilduringthestudyperiod.Themeanconcentrationsinatmospherewereallhigherthanthoseinsoil,andthehighestmethaneconcentrationwasfoundinairattheheightof16mwiththelowestconcentrationoccur-ringatthedepthof1.5minsoil.Thevariationsofatmosphericmethaneconcentrationsdidnotshowanyclearpatternbothtemporallyandspatially,althoughtheyexhibitedamoresteady-stablestatethanthoseinsoil.Duringtheseasonalvariations,themethaneconcentrationsatdifferentdepthsinsoilweresig-nificantlycorrelated(R2>0.6)witheachothercomparingtotheweakcorrelations(R2<0.2)betweentheatmosphericconcentra-tionsatdifferentheights.Meanmethaneconcentrationsinsoilsignificantlydecreasedwithdepth.Thiswasthecompositiveinfluenceofthedecreasingproductionratesandtheincreasingmethaneoxidationrates,whichwascausedbythedescentsoilmoisturewithdepth.Althoughthemethaneconcentrationsatalldepthsvariedwidelyduringthegrowingseason,theyshowedverydistincttemporalvariationsinthenon-growingseason.Itwasindicatedfromtheliteraturesthatmethaneoxidationrateswerepositivelycorrelatedwithsoiltemperature.Thehighermethaneconcentrationsinsoilduringthewinterweredeter-minedbythelowermethaneoxidationrateswithdecreasingsoiltemperatures,whereasmethaneproductionrateshadnoreactiontothelowertemperature.Relationsbetweenmethanecontributionandotherenvironmentalfactorswerenotdiscussedinthispaperforlackingofdata,whichimpulseustocarryoutfurtherandmoredetailedstudiesinthisuniquearea.

简介：Themainobjectiveofthisstudyistosimulatethepotentialvegetationtypesonthebasisofenvironmentalparameters.ThepapertookBarkamCountyinamountainousregionoftheEasternTibetanPlateauasthestudyarea.Thevegetationdistributionwasmappedin1994and2007basedonTMremotesensingimagesbyobject-orientedinterpretationmethod.Weoverlaidthetwomapstofindoutthevegetationpatcheswhichhavenotchanged,andtookthemasstabletypes.Fiftypercentofthestablepatcheswererandomlysampledtooperatethelogisticregressionwithrelatedenvironmentalparameters;otherswereusedastestdataofsimulatedresults.Sevenenvironmentalparametersweremapped,includingelevation,slope,aspect,surfacecurvature,solarradiation,temperatureandprecipitation,basedonDEMdataandmeteorologicalsitedatabyGIStechnology.Therelationshipbetweenthespatialdistributionofvegetationandenvironmentalvariableswerequantifiedbylogisticregression.Thedistributionprobabilitiesofeachvegetationtypewerecalculated.Finally,thespatialdistributionofpotentialvegetationwassimulated.Thisresearchcanprovideascientificbasisforvegetationrestorationandecologicalconstructioninthisarea.