1、 12 5 2010 M 10 o S JOURNAL OF GEOnullINFORM ATION SC IENCEVolnull12, Nonull 5Octnull, 2010l : 2010- 02- 24; : 2010- 06- 12. “ : S E 1 S “ ( 40971236);S E $ ? Z 9 null 973null “ ( 2007CB407207); null S E S / 1 null 9 ( 2007BAC15B01)bT e : ( 1979- ), 3 , 2 7 , p V , Z _ * 3 r m bEnullm ai:l liuf lrei
2、s. ac. cn H W s O q * r b WM M Y 刘 峰1, 2,朱阿兴1, 3,裴 韬1,秦承志1,李宝林1( 1. S S S “ d S E L i , 100101;2. S S 3 , 100049; 3. Departm ent of Geography, Un ivers ity ofW isconsinnullM adison, M adison, W I 53706, USA)K 1 : r w ,* / Y $ T m , 4 $ ,i r W 1 “ w r bW b 7 , j : H W s O q * r bW M M Y - H W Y L Y M
3、 , 9 KullachnullLeibler f , b , D ( YX )V U Y X M bxijayijai j V U i l b T 1 , SID na tbnat , 1 e 9 ; T 2 bitb; V 0 k v b ; - H W Y L M , ; 0; ; - H W Y L M H ,; 0b2. 4 1 Q T s r s “ S _ V Q T s ? 8 C r s by N , i u s ,7 u = r L i s 52 7 r MODIS u V Q r W 1 “ s bm 2A U 52 7 r bW s b 250 m # 250 m ,
4、r E | v r 1 z V b t r L i s (r ) 1 F “ r s 9 F 7 9 F b7 O , F ( 4a5a6 7) , 9 V C B + b8 , M - $ H q / , r u , V Q T A U A s ; M r u , V Q T A M b7 O , r M , Q T t M by N , V , YV H W s O q * 4 | V Q ? r u = r s b4 u , 4 a$ H q ,Y r H q bW x , 4 r M Y r bW M b u , 4 Z E , F r u , Q 9 A ; r M , Q 9 M
5、b B ? C r M 7 ? 4 $ , 4 u r w bL = , F 2. Graduate University of Chinese A cademy of Sciences, B eijing 100049, China;3. D epartm ent of Geography, Un iversity of W isconsinnullM adison, M adison, W I 53706, USA )Abstract: Soil is them ost basic and m ost im portant resource for the living and devel
6、opm ent of hum an beings. A tpresent, there is an urgent need to obtain accurate soil spatial inform ation for environm entalmodeling and agriculnulltural app lications. The theory of soilnulllandscape relationsh ips relates difficultnulltonullm easure soil in form ation, w hich innullcludes soil ty
7、pe and properties, w ith som e easynulltonullob tain soilnullform ing environm ental factors. Thism akes it possibleto infer soil spatial variations from the easynulltonullob tain environm ental factors. Satellite rem ote sensing techniquesgenerally serve as auxiliary tools for researches on soil sp
8、atial in form ation extraction. The techniques can be used toacquire landform and vegetation data. Then, based on their relationsh ipsw ith soil conditions, soil spatial variationcan be derived. In areasw ith m ed ian to strong landform gradien,t easynulltonullm easure landform and vegetation factor
9、shave p layed an important role in soilm app ing. H ow ever, for the p lains and topograph ically gently undulating arenullas, where spatial variation of soilproperties such as soil texture usually cannot be effectively ind icated by the landnullform and vegetation, a scientific prob lem is how to r
10、eveal spatialvariation of soil properties. This paper presen ts anapproach to using land surface dynam ic feedback captured by h igh temporal resolution rem ote sensing over a shortperiod imm ed iately after am ajor rain to d istingu ish soil texture over such areas. The results of this study show e
11、dthat land surface dynam ic feedbacks captured by h igh tem poral resolution rem ote sensing can be used to identifyspatial variation of soil texture. Locations experiencing d ifferent soil texture have obviously d ifferent dynam ic feednullbacks, and them ore sim ilar the soil texture betw een loca
12、tions the more sim ilar their dynam ic feedbacks. The findnullings are very encouraging. The work reported in this paper provides an effective solu tion to the d ifficulty of identifnullying soil spatial variation over low relief areas. It exh ib its the potential of high tem poral rem ote sensing on the exnullaction of soil inform ation.Keywords: high tem poral resolution; rem ote sensing; soil texture; spatial variation740 o S 2010 M
