ArticlePDF Available

Abstract and Figures

Because of the limited spatiotemporal resolutions in vegetation index (VI) products, land surface phenology (LSP) results may not well capture ground-based phenological changes. This is likely the result of the mixed pixel effect: (1) a pixel in VI products may contain an unknown composition of vegetation species or land cover types; and (2) these species differ in their sensitivity to climatic variations. The mixed pixel effect has induced inconsistent findings in LSP with in situ observations of spring phenology. To this end, this study has designed a series of simulation experiments to initiate the methodological exploration of how the green-up date (GUD) of a mixed pixel could be altered by the endmember GUDs and different non-GUD variables, including the endmember composition, minimum and maximum normalized difference vegetation index (NDVI), and the length of the growth period. The study has also compared the sensitivity of two generally adopted GUD identification methods, the relative threshold method and the curvature method (also known as the inflection-point method). The simulations with two endmembers show that even if there is no change in the endmember GUDs, the GUD of the mixed pixel could be substantially altered by the changes in non-GUD variables. In addition, the study has also developed a simulation toolkit for the GUD identification with cases of three or more endmembers. The results of the study provide insights into effective strategies for analyzing spring phenology using VI products: the mixed pixel effect can be alleviated by selecting pixels that are relatively stable in the land cover or species composition. This simulation study calls for in situ phenological observations to validate the LSP, such as conducting climate-controlled experiments on few mixed species at a small spatial scale. The paper also argues for the necessity of isolating GUD trends caused by non-phenological changes in the study of spring phenology.
Content may be subject to copyright.
,;CEJ;-;DI?D=E<DL?HEDC;DJ NNN  NNNNNN
.>; C?N;: F?N;B ;W;9J ?D B7D: IKH<79; F>;DEBE=O  I?CKB7J?ED IJK:O
2?7D= >;DX7X X8 7M;? 17D=X7 $?D >;DX7X X ED= 17D=X7 '?7E=;D ->;DX9
7State Key Laboratory of Earth Surface Processes and Resource Ecology, Beijing Normal University, Beijing 100875, China
8Department of Emergency Management, Arkansas Tech University, Russellville, AR 72801, USA
9Key Laboratory of Alpine Ecology and Biodiversity, Institute of Tibetan Plateau Research, CAS Center for Excellence in Tibetan Plateau Earth Sciences, Beijing 100101, China
,.#& #( )
!H;;DKF :7J; !/
&7D: IKH<79; F>;DEBE=O
'?N;: F?N;B ;<<;9J
;97KI; E< J>; B?C?J;: IF7J?EJ;CFEH7B H;IEBKJ?EDI ?D L;=;J7J?ED ?D:;N 0# FHE:K9JI B7D: IKH<79; F>;DEBE=O &-*
H;IKBJI C7O DEJ M;BB 97FJKH; =HEKD:87I;: F>;DEBE=?97B 9>7D=;I .>?I ?I B?A;BO J>; H;IKBJ E< J>; C?N;: F?N;B ;<
<;9J  7 F?N;B ?D 0# FHE:K9JI C7O 9EDJ7?D 7D KDADEMD 9ECFEI?J?ED E< L;=;J7J?ED IF;9?;I EH B7D: 9EL;H JOF;I
7D:  J>;I; IF;9?;I :?<<;H ?D J>;?H I;DI?J?L?JO JE 9B?C7J?9 L7H?7J?EDI .>; C?N;: F?N;B ;<<;9J >7I ?D:K9;: ?D9ED
I?IJ;DJ UD:?D=I ?D &-* M?J> ?D I?JK E8I;HL7J?EDI E< IFH?D= F>;DEBE=O .E J>?I ;D: J>?I IJK:O >7I :;I?=D;: 7 I;H?;I
E< I?CKB7J?ED ;NF;H?C;DJI JE ?D?J?7J; J>; C;J>E:EBE=?97B ;NFBEH7J?ED E< >EM J>; =H;;DKF :7J; !/ E< 7 C?N;:
F?N;B 9EKB: 8; 7BJ;H;: 8O J>; ;D:C;C8;H !/I 7D: :?<<;H;DJ DED!/ L7H?78B;I ?D9BK:?D= J>; ;D:C;C8;H 9EC
FEI?J?ED C?D?CKC 7D: C7N?CKC DEHC7B?P;: :?<<;H;D9; L;=;J7J?ED ?D:;N (0# 7D: J>; B;D=J> E< J>; =HEMJ>
F;H?E: .>; IJK:O >7I 7BIE 9ECF7H;: J>; I;DI?J?L?JO E< JME =;D;H7BBO 7:EFJ;: !/ ?:;DJ?U97J?ED C;J>E:I J>; H;B
7J?L; J>H;I>EB: C;J>E: 7D: J>; 9KHL7JKH; C;J>E: 7BIE ADEMD 7I J>; ?DV;9J?EDFE?DJ C;J>E: .>; I?CKB7J?EDI
M?J> JME ;D:C;C8;HI I>EM J>7J ;L;D ?< J>;H; ?I DE 9>7D=; ?D J>; ;D:C;C8;H !/I J>; !/ E< J>; C?N;: F?N;B
9EKB: 8; IK8IJ7DJ?7BBO 7BJ;H;: 8O J>; 9>7D=;I ?D DED!/ L7H?78B;I #D 7::?J?ED J>; IJK:O >7I 7BIE :;L;BEF;: 7
I?CKB7J?ED JEEBA?J <EH J>; !/ ?:;DJ?U97J?ED M?J> 97I;I E< J>H;; EH CEH; ;D:C;C8;HI .>; H;IKBJI E< J>; IJK:O
FHEL?:; ?DI?=>JI ?DJE ;<<;9J?L; IJH7J;=?;I <EH 7D7BOP?D= IFH?D= F>;DEBE=O KI?D= 0# FHE:K9JI J>; C?N;: F?N;B ;<<;9J
97D 8; 7BB;L?7J;: 8O I;B;9J?D= F?N;BI J>7J 7H; H;B7J?L;BO IJ78B; ?D J>; B7D: 9EL;H EH IF;9?;I 9ECFEI?J?ED .>?I I?CK
B7J?ED IJK:O 97BBI <EH ?D I?JK F>;DEBE=?97B E8I;HL7J?EDI JE L7B?:7J; J>; &-* IK9> 7I 9ED:K9J?D= 9B?C7J;9EDJHEBB;:
;NF;H?C;DJI ED <;M C?N;: IF;9?;I 7J 7 IC7BB IF7J?7B I97B; .>; F7F;H 7BIE 7H=K;I <EH J>; D;9;II?JO E< ?IEB7J?D=
!/ JH;D:I 97KI;: 8O DEDF>;DEBE=?97B 9>7D=;I ?D J>; IJK:O E< IFH?D= F>;DEBE=O
1. Introduction
&7D: IKH<79; F>;DEBE=O &-* IJK:?;I J>; I;7IED7B?JO E< L;=;J7J;:
B7D: IKH<79;I 8O H;CEJ;BO I;DI;: ?C7=;H?;I 1>?J; ;J 7B  &-*
L?;M;: 7I 7 BED=J;HC =BE87BBOI;DI;: B7D:I97F; <;7JKH; ?I 78B; JE EL;H
9EC; J>; B?C?J;: IF7J?EJ;CFEH7B 9EL;H7=; ?D>;H;DJ ?D FBEJ87I;: ?DI?JK
E8I;HL7J?EDI ;= >KC7D E8I;HL7J?EDI 97DEFO 97C;H7I VKN JEM;HI E<
J>; ;9EIOIJ;C I 7 H;IKBJ &-* I;HL;I 7D ;<<;9J?L; HEB; ?D 9ECFB;C;DJ
,;;: ;J 7B  4>7E 7D: -9>M7HJP  #J FHEL?:;I 9EHHE8EH7J?D=
;L?:;D9; <EH L7B?:7J?D= ;9EBE=?97B H;IFEDI;I JE >?IJEH?97B 9B?C7J; 9>7D=;
I9;D7H?EI 7:;9A ;J 7B 
/I?D= L;=;J7J?ED ?D:?9;I 0#I DEJ78BO J>; DEHC7B?P;: :?<<;H;D9;
L;=;J7J?ED ?D:;N (0# 7D: J>; ;D>7D9;: L;=;J7J?ED ?D:;N 0# &-*
IJK:?;I 97D :;H?L; F>;DEBE=?97B C;JH?9I 7I ?D:?97J?EDI E< ;9EIOIJ;C :O
D7C?9I 4>7D= ;J 7B  .>; =H;;DKF :7J; !/ 7BIE ADEMD 7I
J>; IJ7HJE<I;7IED 5-)-6 7I 7D ?CFEHJ7DJ F>;DEBE=?97B C;JH?9 9>7H79
J;H?P;I J>; EDI;J E< C;7IKH78B; F>EJEIODJ>;J?9 79J?L?J?;I EH J>; J?C?D=
E< IFH?D= 7HH?L7BI ;IF?J; J>; FHEB?<;H7J?D= 7FFB?97J?EDI E< J>; !/ 7I
7 IODEFJ?9 L7H?78B; ?D H;I;7H9> ED 9B?C7J; 9>7D=; HEMD ;J 7B 
->;D ;J 7B  -Jö9AB? 7D: 0?:7B;  7H=KC;DJI 7HEI; 78EKJ J>;
H;B?78?B?JO E< J>; &-* ?D H;FB?97J?D= =HEKD:87I;: F>;DEBE=?97B ;L;DJI
7D: J;IJ?D= J>; 9B?C7J?9 I;DI?J?L?JO E< L;=;J7J?ED 1>?J; ;J 7B 
IK9> 7I J>; 7H=K78B; 7:L7D9; E< J>; !/ ?D J>; .?8;J7D *B7J;7K 7I 7 H;
IFEDI; JE 9B?C7J; 9>7D=; ->;D ;J 7B  4>7D= ;J 7B 
#D 7::?J?ED JE 7JCEIF>;H?9 ?DJ;H<;H;D9; ;= 9BEK:I 7;HEIEBI 7D:
7B=EH?J>C?9 9EHH;9J?EDI J>7J 9EDJ7C?D7J; EH CE:?<O J>; JHK; IF;9JH7B
?D<EHC7J?ED >B ;J 7B  JME ?CFEHJ7DJ =;E=H7F>?9 KD9;HJ7?D
J?;I C7O E8<KI97J; J>; ?:;DJ?U97J?ED E< J>; !/ 8;97KI; E< J>; 9E7HI;
IF7J?EJ;CFEH7B H;IEBKJ?EDI ?D 0# FHE:K9JI ;= !#''- ')#- -*).
0!. .>; UHIJ IEKH9; E< KD9;HJ7?DJO F;HJ7?DI JE J>; J;CFEH7B C?I
Email address: 9>;D@?D8DK;:K9D $ >;D
,;9;?L;: $KD;,;9;?L;:?DH;L?I;:<EHC '7H9>99;FJ;: FH?B
L7?B78B;EDB?D; NNN
 R 
X. Chen et al. Remote Sensing of Environment xxx (2018) xxx-xxx
C7J9> E< H;CEJ;BO I;DI;: :7J7 M?J> =HEKD: E8I;HL7J?EDI EH ;N7CFB;
J>; CE:;H7J; H;IEBKJ?ED ?C7=?D= IF;9JHEH7:?EC;J;H ')#- 0# FHE:
K9JI 97DDEJ 9ECFB;J;BO C7J9> J>; =HEKD: :7J;I 8;97KI; J>; L;=;J7
J?ED =HEMJ> 8;JM;;D J>; =H;;DKF 7D: J>; C7JKH?JO ?I 7 :OD7C?9 7D:
H7F?: FHE9;II ;= KI?D= 7D  EH :7O ')#- FHE:K9JI JE :;J;9J 7
:7O =HEMJ> F;H?E: >B ;J 7B  .>; I;9ED: IEKH9; E< KD
9;HJ7?DJO ?I J>; C?N;: F?N;B ;<<;9J 7BIE ADEMD 7I J>; FE?DJ LI F?N;B
FHE8B;C ,;;: ;J 7B  1>?J; ;J 7B  .>; C?N;: F?N;B ;<<;9J
7H?I;I 8;97KI; J>7J  7 F?N;B ?D 0# FHE:K9JI C7O 9EDJ7?D 7D KDADEMD
9ECFEI?J?ED E< L;=;J7J?ED IF;9?;I EH B7D: 9EL;H JOF;I 8EJ> 97D 8; H;
=7H:;: 7I ;D:C;C8;HI ?D IF;9JH7B C?NJKH; CE:;BI 7D: J>7J  J>;I;
IF;9?;I :?<<;H ?D J>;?H I;DI?J?L?JO JE 9B?C7J?9 L7H?7J?EDI ?D J;HCI E< J>;
!/I 7:;9A ;J 7B  K9>;C?D ;J 7B  -9>M7HJP 7D: ,;;:
 4>7D= ;J 7B  #D 7::?J?ED J>; =H;;DD;II E< KD:;HIJEHO F>;
DEBE=O <KHJ>;H 9ECFB?97J;I J>; :;J;9J?D= E< EL;HIJEHO I?=D7B >B ;J 7B
 .H;C8B7O 7D: &7HE9GK;  I J>; F?N;B87I;: !/ JH;D: 97D
8; ?D:K9;: 8O J>; 9>7D=; E< 7 I?D=B; L7H?78B; EH 7 9EC8?D7J?ED E< L7H?
78B;I ;= IF;9?;I 9ECFEI?J?EDI IF;9?;I !/I ;D97FIKB7J;: ?D 7 C?N;:
F?N;B ?J ?I <7H <HEC 9B;7H 78EKJ J>; FH;9?I; C;9>7D?IC J>7J :?9J7J;I J>;
J>; ;<<;9J IK9> 7I 9ECF7H?D= J>; F>;DEBE=?97B :;J;9J?EDI 8;JM;;D ?C
7=;I E< :?<<;H;DJ H;IEBKJ?EDI 4>7D= ;J 7B  ;N9;FJ <;M 97I;I M>;H;
C?N;: IF;9?;I M;H; ;N7C?D;: I;F7H7J;BO 7:;9A ;J 7B  K9>;C?D
;J 7B 
BJ>EK=> J>; C?N;: F?N;B ;<<;9J ?D J>; &-* >7I H;9;?L;: ?D9H;7I?D=
7JJ;DJ?ED ?J ?I KD9B;7H 7I JE M>7J ;NJ;DJ J>; ;<<;9J ?I 79JK7BBO I>7F?D=
J>; IKH<79; H;V;9J7D9; 7D: ;L;DJK7BBO 9EDJH?8KJ?D= JE J>; E8I;HL;: !/
I>?<J /D:;H J>; 9EDJ;NJ E< 9B?C7J; 9>7D=; J>; ?DJ;H7DDK7B :OD7C?9I E<
J>; &-* ?I DEJ D;9;II7H?BO J>; H;IKBJ E< I>?<J;: F>;DEBE=?97B F7JJ;HDI 8KJ
9EKB: 8; F7HJ?7BBO 7JJH?8KJ;: JE J>; 7BJ;H7J?ED E< J>; B7D: 9EL;H IK9> 7I
J>; ?DJHE:K9J?ED E< 7 D;M IF;9?;I EH J>; ;NF7DI?ED E< 7=H?9KBJKH7B B7D:I
";BC7D  .E J>?I ;D: J>?I F7F;H >7I ;NFBEH;: >EM J>; !/ E<
7 C?N;: F?N;B 9EKB: 8; C;9>7D?IJ?97BBO 7BJ;H;: 8O 8EJ> J>; ;D:C;C8;H
!/I 7D: :?<<;H;DJ DED!/ L7H?78B;I ?D 7D 7DDK7B :;L;BEFC;DJ 9O9B;
2. Method
2.1. Simulation of annual NDVI temporal pro9le
.>; 7DDK7B (0# :OD7C?9I H;FH;I;DJ?D= J>; =HEMJ> EH I;D;I9;D9;
F;H?E: E< 7 FB7DJ IF;9?;I m?D 7 :;L;BEFC;DJ 9O9B; 97D 8; :;I9H?8;: 8O
7 BE=?IJ?9 CE:;B 7I I>EMD ?D G  ,7JAEMIAO  4>7D= ;J 7B
M>;H; t?I J>; :7O E< J>; O;7H )3 a7D: b7H; F7H7C;J;HI 7IIE9?7J;:
M?J> J>; J?C?D= 7D: 9>7D=; H7J; b <EH J>; =HEMJ> F;H?E: 7D: 8
<EH J>; I;D;I9;D9; F;H?E: E< J>; (0# JH7@;9JEHO c?I J>; 7CFB?JK:; E<
J>; JH7@;9JEHO 7D: d?I J>; C?D?CKC (0# L7BK; (0#XC?D #D J>?I H;
IF;9J J>; L7BK; E< 9: O?;B:I J>; C7N?CKC (0# L7BK; (0#XC7N 7J
J>; J?C?D= E< <KBB C7JKH?JO
1; IJ7HJ;: J>; 7D7BOI?I M?J> 7 I?CKB7J;: C?N;: F?N;B 9ECFEI;: E<
JME ;D:C;C8;HI ;D:C;C8;H  8;BED=I JE 7 :;9?:KEKI <EH;IJ 7D: ;D:
C;C8;H  8;BED=I JE 7 9HEFB7D: .>; 7DDK7B (0# FHEUB;I E< J>;I; JME
;D:C;C8;HI M;H; :;H?L;: <HEC J>; ')#- 0# FHE:K9JI :7O '0
C ')+ EBB;9J?ED  "K;J; ;J 7B  87I;: ED J>; O;7H
 J;CFEH7B FHEUB;I E< ;D:C;C8;HI ?D JME I;B;9J;: B7D: F7H9;BI
?D (EHJ>;7IJ >?D7 .>; FHE9;II E< J>; :7J7 9EBB;9J?ED ?I ?D9BK:;: ?D J>;
-KFFB;C;DJ7HO 7J7 .>; 7L;H7=;: (0# FHEUB;I E< J>; JME ;D:C;C
8;HI 7H; I>EMD ?D ?= 7 .>; !/I E< J>; ;D:C;C8;HI M;H; ?:;DJ?U;:
7J J>; ?DV;9J?ED FE?DJ E< J>; BE=?IJ?9 9KHL; M>;D J>; H7J; E< 9>7D=; ?D
9KHL7JKH; H;79>;: J>; UHIJ BE97B C7N?CKC 4>7D= ;J 7B  1; J>;D
=;D;H7J;: 7 C?N;: F?N;B 87I;: ED IF;9?U9 9ECFEI?J?EDI E< J>; JME ;D:
C;C8;HI .>; (0# L7BK; E< J>; C?N;: F?N;B (0#XC?N M7I =;D;H7J;:
7I 7 B?D;7H 9EC8?D7J?ED E< J>; (0# L7BK;I E< ;D:C;C8;HI M;?=>J;: 8O
J>;?H 9EDJH?8KJ?ED <79JEHI :7CI ;J 7B  7I I>EMD ?D G  7D:
?= 8
M>;H; (0#XC :;DEJ;I J>; (0# L7BK; E< ;D:C;C8;H C 9ECFEI;: E<
Fig. 1. #BBKIJH7J?ED E< ?:;DJ?<O?D= J>; !/XC?N M?J> JME ;D:C;C8;HI ?D 7 I?CKB7J;: C?N;: F?N;B 7 (0# J;CFEH7B FHEUB;I E< J>; JME ;D:C;C8;HI 8 (0#XC?N M?J> :?<<;H;DJ ;D:C;C8;H
9ECFEI?J?EDI 9 !/XC?N 8O J>; JME ?:;DJ?U97J?ED C;J>E:I
X. Chen et al. Remote Sensing of Environment xxx (2018) xxx-xxx
L;=;J7J?ED 7D: IE?B 879A=HEKD: fXC :;DEJ;I J>; 9EDJH?8KJ?ED <79JEH E<
;D:C;C8;H C ?D J>; C?N;: F?N;B 7D: ' ?I J>; DKC8;H E< ;D:C;C8;HI
?D J>; C?N;: F?N;B ' ?D J>?I IJK:O
.>; EH?=?D7B (0# L7BK;I E< J>; ;D:C;C8;HI M;H; ;NJH79J;: <HEC
7 ')#- F?N;B ?D M>?9> 8EJ> L;=;J7J?ED 7D: IE?B 879A=HEKD: I?=D7BI
M;H; ?D9BK:;: .>; ?D:;F;D:;DJ 9EDJH?8KJ?ED E< IE?B JE 7 C?N;: F?N;B
J>KI 97DDEJ 8; 9EDI?:;H;: ?D G  'EH;EL;H 7I J>; (0# ?I 7D ?DJ;
=H7J;: I?=D7B :?9J7J;: 8O 8EJ> J>; =H;;D L;=;J7J?ED <H79J?ED !0  7D:
J>; B;7< 7H;7 ?D:;N &# !7E ;J 7B  fXC M7I KI;: JE H;FH;I;DJ J>;
8?B7J;H7B 9EDJH?8KJ?EDI <HEC 8EJ> J>; !0 7D: J>; &# H7J>;H J>7D J>;
L;=;J7J?ED <H79J?ED F;H I; 'EH; ?CFEHJ7DJBO 7BJ>EK=> J>; B?D;7H IF;9
JH7B C?NJKH; CE:;B M7I IJH?9JBO L7B?: <EH J>; EH?=?D7B H;V;9J7D9; L7BK;I
;N?IJ?D= IJK:?;I :;CEDIJH7J;: J>7J J>?I B?D;7H C?NJKH; CE:;B MEKB: ?D
JHE:K9; EDBO C?DEH 7D: D;=B?=?8B; ;HHEHI M>;D H;FB79?D= J>; H;V;9J7D9;
8O J>; DEDB?D;7HBO JH7DI?J?ED;: (0# C7D ;J 7B  %;H:?B;I 7D:
!HED:ED7  .>?I ;L?:;D9; ;DIKH;: 7D 799;FJ78B; B;L;B E< 799KH79O
<EH :;H?L?D= J>; 7==H;=7J; (0# E< J>; I?CKB7J;: C?N;: F?N;B 7I;: ED
G  J>; 7DDK7B (0#XC?N J;CFEH7B FHEUB;I M?J> :?<<;H;DJ fXC M;H; =;D
;H7J;:  ?= 8
2.2. Detection of GUD⁠mix
:;J;9J?D= J>; IFH?D= F>;DEBE=O ?D J;HCI E< J>; !/XC?N  J>; H;B7J?L;
J>H;I>EB: C;J>E: $öDIIED 7D: ABKD:>  1>?J; ;J 7B  3K ;J
7B  7D:  J>; 9KHL7JKH; C;J>E: 7BIE 97BB;: J>; ?DV;9J?EDFE?DJ
C;J>E: 4>7D= ;J 7B  4>7D= 7D: !EB:8;H=  M>?9> M7I
KI;: JE FHE:K9; J>; ')#- F>;DEBE=?97B FHE:K9J '+ 4>7D= ;J
7B  .>;I; JME C;J>E:I M;H; 9ECF7H;: 7D: J;IJ;: <EH J>;?H I;D
I?J?L?JO JE J>; C?N;: F?N;B ;<<;9J #D J>; H;B7J?L; J>H;I>EB: C;J>E: J>;
!/ E< J>; C?N;: F?N;B !/XC?N M7I ?:;DJ?U;: 7I J>; :7O M>;D J>;
(0#XC?N H;79>;: 7 IF;9?U9 F;H9;DJ7=; ?;  E< ?JI 7DDK7B 7CFB?
JK:; 1>?J; ;J 7B  EH J>; 9KHL7JKH; C;J>E: J>; I?=CE?:I>7F;:
BE=?IJ?9 <KD9J?ED M7I UJJ;: JE J>; (0#XC?N :7J7 <EBBEM;: 8O J>; ?:;DJ?U
97J?ED E< J>; !/XC?N 7I J>; :7O M>;D J>; H7J; E< 9>7D=; ?D J>; (0#XC?N
9KHL7JKH; H;79>;: ?JI UHIJ BE97B C7N?CKC  ?= 9
2.3. Simulation experiments
.>; H7J?ED7B; E< J>; I?CKB7J?ED 7FFHE79> M7I JE ;L7BK7J; J>; I;DI?
J?L?JO E< J>; !/XC?N JE <EKH DED!/ L7H?78B;I fXC (0#XC7N (0#XC?D
7D: J>; =HEMJ> F;H?E: 8;JM;;D J>; !/ 7D: J>; C7JKH?JO :7J; 7D: J>;
;D:C;C8;H !/I KD:;H :?<<;H;DJ I9;D7H?EI .78B; 7D: ?=  79>
J;IJ L7H?78B; M7I I?CKB7J;:  J?C;I 9EHH;IFED:?D= JE JM;DJO O;7HI E<
F>;DEBE=?97B 9>7D=;I 8O 7:@KIJ?D= J>; L7H?78B; 7J 7D ;GK7B ?D9H;C;DJ
;= ?D ?= 8 J>; (0#XC7N E< ;D:C;C8;H  :;9H;7I;I <HEC  JE 
7J 7D ?D9H;C;DJ E<  M>;H; EDBO UL; A;O L7H?78B;I M;H; KI;: <EH
8;JJ;H L?IK7B?P7J?ED .>; H7J?ED7B; E< 9>EEI?D= J>; H7D=; E< J>; J;IJ M7I
87I;: ED J>; IJ7J?IJ?9I E< J>; L7H?78B; ?D J>; O;7H ')#- 0# FHE:K9J
I;; J>; -KFFB;C;DJ7HO 7J7 <EH :;J7?BI
3. Results
-9;D7H?E # 9>7D=;I J>; fX7 <HEC  JE  7D: J>KI fX8 <HEC  JE 
7J 7D ?D9H;C;DJ E<  1?J> J>; JME EH?=?D7B (0# J;CFEH7B FHEUB;I
KD9>7D=;: ?D9H;7I?D= J>; fX7 J>; ;7HBO !/ ;D:C;C8;H 9EDI?:;H78BO
7:L7D9;I J>; !/XC?N 7I I>EMD ?D ?=  #D J>; EJ>;H H;IF;9J ?D9H;7I?D=
J>; fX8 J>; B7J; !/ ;D:C;C8;H :;B7OI J>; !/XC?N JE 7 9EDI?:;H78B;
;NJ;DJ DEJ L?IK7B?P;: ?D J>; 7HJ?9B; .>?I I9;D7H?E FHEL?:;I 9EHHE8EH7J
?D= ;L?:;D9; E< J>; C?N;: F?N;B ;<<;9J :?I9KII;: ?D EJ>;H B?J;H7JKH; J>;
!/XC?N 97D 8; 9EDI?:;H78BO 7<<;9J;: 8O 9>7D=;I ?D <H79J?EDI E< J>; ;D:
C;C8;HI M?J>?D 7 F?N;B ;L;D ?< DE F>;DEBE=?97B 9>7D=;I >7L; E99KHH;:
";BC7D  4>7D= ;J 7B  .>KI J>; H;CEJ;BO I;DI;: !/XC?N
I>?<JI ED J>; F?N;B B;L;B 9EKB: 8; 7 H;IKBJ E< 9>7D=;I ?D J>; B7D: 9EL;H EH
Table 1
-9;D7H?E 07H?78B; N7CFB; (0# Δ!/XC?N
## '7N?CKC
&?7D: EN
## '?D?CKC
## ,7J;E<
### D:C;C8;H
?= I>EMI J>; H;IKBJI <EH -9;D7H?EI ## ## 7D: ## 8O CE:?<O?D=
J>; (0# FHEUB; E< ED; ;D:C;C8;H M>?B; C7?DJ7?D?D= 7D ;GK7B 9EDJH?
8KJ?ED <79JEH fX7 fX8  D ?CFEHJ7DJ FH;C?I; <EH J>;I; I9;D7H?EI ?I
J>7J J>; ;D:C;C8;H !/I H;C7?D KD9>7D=;: IE 7DO 9>7D=; C7:; JE
J>; ;D:C;C8;H (0# FHEUB; IK9> 7I J>; (0#XC7N  ?= 7 7D: 8 J>;
(0#XC?D  ?= 9 7D: : 7D: J>; B;D=J> E< J>; =HEMJ> F;H?E:  ?= ;
7D: < C7O 9ED<EKD: J>; ?:;DJ?U97J?ED E< J>; !/XC?N #D -9;D7H?E ##
 ?= 7 7D: 8 M>;D J>; (0#XC7N E< ;D:C;C8;H  :;9H;7I;I <HEC 
JE  J>; !/XC?N ?I :;B7O;: 8O 78EKJ :7OI C7N QΔ!/XC?NQ #D
J>; EJ>;H H;IF;9J :;9H;7I?D= J>; (0#XC7N E< ;D:C;C8;H  7:L7D9;I J>;
!/XC?N #D -9;D7H?E ##  ?= 9 7D: : ?D9H;7I?D= J>; (0#XC?D E< ;D:
C;C8;H  :;B7OI J>; !/XC?N 7D: J>?I :;B7O ;<<;9J C7N QΔ!/XC?NQ 
?I I?C?B7H JE :;9H;7I?D= J>; (0#XC7N 8KJ ?I 9EDI?:;H78BO B;II IK8IJ7DJ?7B
J>7D 9>7D=?D= J>; fXC C7N QΔ!/XC?NQ #D 7::?J?ED J>; !/XC?N ?D
-9;D7H?E ##  ?= ; 7D: < ?I 7BIE I>?<J;: M>;D J>; =HEMJ> F;H?E: E<
ED; ;D:C;C8;H ?I ;NJ;D:;: M?J>EKJ J>; ;D:C;C8;H !/ I>?<J C7N
QΔ!/XC?NQ #D 7BB J>;I; J;IJI J>; 9KHL7JKH; C;J>E: ?I CEH; I;DI?
J?L; JE J>; 9>7D=;I J>7D J>; H;B7J?L; J>H;I>EB: C;J>E: 1; 7BIE 9>7D=;:
J>; 9EDJH?8KJ?ED <79JEHI E< ;D:C;C8;HI  7D:  fXC  M>?B;
J>; !/XC?N JH;D:I 7H; 9EDI?IJ;DJ M?J> J>; 97I; E< fXC  8KJ JE :?<<;H
;DJ 7:L7D9; EH :;B7O :;=H;;I H;IKBJI 7H; DEJ I>EMD ?D J>; F7F;H
?= I>EMI J>; H;IKBJI E< J>; Δ!/XC?N ?D -9;D7H?E ### 8O 9>7D=
?D= J>; ;D:C;C8;H !/I I J>; H;IKBJI ?BBKIJH7J; 7:L7D9?D= J>; !/
E< ;D:C;C8;H  9EDI?:;H78BO 7:L7D9;I J>; !/XC?N M>?B; :;B7O?D= J>;
!/ E< ;D:C;C8;H  >7I 7D EFFEI?J; 8KJ B;II I?=D?<?97DJ ;<<;9J .>?I
H;IKBJ ?I 7BIE 9EHHE8EH7J;: 8O 7 H;9;DJ UD:?D= J>7J J>; !/XC?N :;J;9
J?ED ?I :?9J7J;: FH?C7H?BO 8O J>; ;7HBO !/ F?N;BI H7J>;H J>7D J>; B7J;
!/ F?N;BI 4>7D= ;J 7B  #D 7::?J?ED M; F;H<EHC;: J>; I;DI?
J?L?JO J;IJ <EH J>; 9EDJH?8KJ?ED <79JEH fXC  M>;H; fXC 97D I?=
D?<?97DJBO EWI;J J>; :;=H;; JE M>?9> J>; !/XC?N ?I 7:L7D9;: EH :;
B7O;: EH ;N7CFB; ?D J>; 97I; E< :;B7O?D= J>; !/ E< ;D:C;C8;H
X. Chen et al. Remote Sensing of Environment xxx (2018) xxx-xxx
Fig. 2. #BBKIJH7J?ED E< J>; I?CKB7J?ED I9;D7H?EI 7 9>7D=?D= fX7 <HEC  JE  7D: J>; (0#I E< 8EJ> ;D:C;C8;HI H;C7?D KD9>7D=;: 8 :;9H;7I?D= J>; (0#XC7N E< ;D:C;C8;H  <HEC
 JE  9 :;9H;7I?D= J>; (0#XC7N E< ;D:C;C8;H  <HEC  JE  : ?D9H;7I?D= J>; (0#XC?D E< ;D:C;C8;H  <HEC  JE  ; ?D9H;7I?D= J>; (0#XC?D E< ;D:C;C8;H  <HEC 
JE  < ;NJ;D:?D= J>; =HEMJ> F;H?E: E< ;D:C;C8;H  <HEC  JE :7OI 8O :;B7O?D= J>; C7JKH?JO :7J; = ;NJ;D:?D= J>; =HEMJ> F;H?E: E< ;D:C;C8;H  <HEC  JE :7OI 8O :;B7O?D=
J>; C7JKH?JO :7J; > 7:L7D9?D= J>; !/ E< ;D:C;C8;H  8O :7OI 7D: ? :;B7O?D= J>; !/ E< ;D:C;C8;H  8O :7OI
8O 7 C7N?CKC E< :7OI J>; !/XC?N ?I D;7HBO KD7<<;9J;: fX8  EH
 ?= 9 .>?I H;IKBJ ?I H;7IED78B; ;L?:;D9; E< J>; C?N;: F?N;B ;<
<;9J J>; !/ 9>7D=; ED J>; F?N;B B;L;B C7O DEJ 8; 7:;GK7J;BO ?:;DJ?U;:
M>;D KD9;HJ7?DJ?;I ;N?IJ ?D 8EJ> J>; fXC 7D: J>; IF;9?;I !/ JH;D:I
4. Discussion and conclusions
.>?I F7F;H :;BL;I ?DJE J>; C;J>E:EBE=?97B KD9;HJ7?DJO E< J>; !/
?:;DJ?U97J?ED 7 FHE8B;C ?D:K9;: 8O J>; C?N;: F?N;B ;<<;9J ?D 0# FHE:
X. Chen et al. Remote Sensing of Environment xxx (2018) xxx-xxx
Fig. 3. >7D=;I ?D !/XC?N Δ!/XC?N KD:;H -9;D7H?E # 8O 7 J>; H;B7J?L; J>H;I>EB: C;J>E: 7D: 8 J>; 9KHL7JKH; C;J>E:
Fig. 4. >7D=;I ?D !/XC?N Δ!/XC?N KD:;H 78 -9;D7H?EI ## 9>7D=?D= J>; (0#XC7N 9: ## 9>7D=?D= J>; (0#XC?D 7D: ;< ## 9>7D=?D= J>; B;D=J> E< J>; =HEMJ> F;H?E:
K9JI BJ>EK=> J>; C?N;: F?N;B ;<<;9J >7I 8;;D H;9E=D?P;: ?D F7IJ B?J;H
7JKH; ;= 7:;9A ;J 7B  ";BC7D  ->;D ;J 7B  J>?I
F?BEJ IJK:O ?I 7CED= J>; UHIJ JE ;N7C?D; J>; I;DI?J?L?JO E< J>; !/ J>7J
C7O 8; 9ED<EKD:;: 8O J>; KD9;HJ7?D 9EDIJ?JK;DJI ?D J>; ?DJ;=H7J?ED E<
IF;9JH7B <;7JKH;I .>HEK=> J>; I;DI?J?L?JO J;IJ KD:;H UL; I9;D7H?EI 7D:
9ECF7H?D= JME !/ ?:;DJ?U97J?ED C;J>E:I J>; I?CKB7J?ED 7FFHE79>
?;I JE ;IJ?C7J; J>; F?N;B87I;: L;=;J7J?ED !/ 9>7D=;I ?D J>; 9EDJ;NJ E<
9B?C7J; 9>7D=;
.>; C7@EH 9EDJH?8KJ?ED E< J>?I H;I;7H9> ?I JE >?=>B?=>J J>; ?D<;H;D
J?7B ;HHEHI 8;JM;;D I7J;BB?J;:;H?L;: &-* 7D: ?D I?JK F>;DEBE=?97B E8
I;HL7J?EDI ED B?C?J;: ?D:?L?:K7BI E< FB7DJ IF;9?;I BJ>EK=> J>; L;=;
I>7F; J>; F>;DEBE=?97B 9O9B; ?D L7H?EKI M7OI ?J H;C7?DI KD9B;7H J>7J
?< IK9> H;IFEDI;I 97D 8; 799KH7J;BO :;J;9J;: 7D: CED?JEH;: 8O I7J;BB?J;
C;7IKH;C;DJI 7:;9A ;J 7B  .>; B79A E< 9EDI?IJ;D9O 8;JM;;D
H;CEJ;BO I;DI;: :7J7I;JI 7D: =HEKD:87I;: E8I;HL7J?EDI 1>?J; ;J 7B
X. Chen et al. Remote Sensing of Environment xxx (2018) xxx-xxx
Fig. 5. >7D=;I ?D !/XC?N Δ!/XC?N M?J> H;IF;9J JE 9>7D=;I ?D !/ Δ!/ E< 7 8 ;D:C;C8;H  7D: 9 : ;D:C;C8;H  KD:;H -9;D7H?E ### M?J> :?<<;H;DJ fXC
 9EKB: 8; 7 H;IKBJ E< JME :?<<;H;DJ IEKH9;I E< ?D<;H;DJ?7B ;HHEHI ?D
:K9;: 8O J>; C?N;: F?N;B ;<<;9J 7I ?BBKIJH7J;: ?D .78B;  ?HIJ J>;H; ?I
7 FEII?8?B?JO J>7J =HEKD:87I;: !/ E8I;HL7J?EDI ED B?C?J;: ?D:?L?:K7BI
E< FB7DJ IF;9?;I C7O DEJ >7L; 9>7D=;: 7I >OFEJ>;I?P;: 8KJ J>; 7==H;
=7J; 0# JH7@;9JEHO C7O C?I?DJ;HFH;J J>; <79J 8;97KI; E< J>; B7D: 9EL;H
9>7D=; EH 7BJ;H;: IF;9?;I 9ECFEI?J?ED -9;D7H?E # EH 9>7D=;I E< EJ>;H
DED!/ L7H?78B;I -9;D7H?E ## .>?I C?I?DJ;HFH;J7J?ED ?I I?C?B7H JE J>;
.OF;  ;HHEH ?D >OFEJ>;I?I J;IJ?D= M>;H; J>; !/ I>?<J ?I ;HHED;EKIBO
F;H9;?L;: -;9ED: ?J ?I 7BIE FEII?8B; J>7J J>; =HEKD:87I;: !/ 9>7D=;
?I E8I;HL;: 8KJ J>; &-* <7?BI JE ?:;DJ?<O J>; JH;D: EH KD:;H;IJ?C7J; J>;
C7=D?JK:; 8;97KI; 9>7D=;I E< L7H?78B;I ?D EFFEI?J; :?H;9J?EDI 97D EW
I;J J>; ;<<;9J -9;D7H?E ### .>?I JOF; E< C?I?DJ;HFH;J7J?ED ;GK?L7B;DJ JE
J>; .OF; ## ;HHEH ?D J>; F7HB7D9; E< >OFEJ>;I?I J;IJ?D= 9EKB: E8<KI97J;
J>; JHK; ?CF79J E< 9B?C7J; 9>7D=;
.>; I?CKB7J?ED H;IKBJI FHEL?:; ?DI?=>JI ?DJE ;<<;9J?L; IJH7J;=?;I <EH
7D7BOP?D= IFH?D= F>;DEBE=O KI?D= 0# FHE:K9JI ?HIJ J>; fXC E< 7D ;D:
C;C8;H ?D 7 C?N;: F?N;B >7I C7HA;: ?DVK;D9;I ED J>; ?:;DJ?U97J?ED
E< J>; !/XC?N 9ECF7H?D= JE EJ>;H L7H?78B;I .>KI M>;D ;NJH79J?D= &-*
C;JH?9I J>; >ECE=;D;?JO E< J>; B7D:I97F; CKIJ 8; ;DIKH;: 8O I;B;9J
?D= F?N;BI J>7J 7H; H;B7J?L;BO IJ78B; ?D J>; B7D: 9EL;H EH IF;9?;I 9ECFE
I?J?ED EL;H J>; J?C; F;H?E: E< E8I;HL7J?ED EH ;N7CFB; 7I I>EMD 8O
K9>;C?D ;J 7B  7D: &?7D= ;J 7B  J>; 9E>;I?ED 8;JM;;D
J>; &-* 7D: J>; =HEKD: E8I;HL7J?ED 97D 8; B7H=;BO ?CFHEL;: ?< J>;
-;9ED: H;9;DJBO IEC; U;B: ;NF;H?C;DJI >7L; H;IEHJ;: JE J>; E8I;HL7
Table 2
#D<;H;DJ?7B ;HHEHI ?D J>; ?:;DJ?U97J?ED E< !/ JH;D: ?D:K9;: 8O J>; C?N;: F?N;B ;<<;9J
!HEKD:87I;:!/JH;D: ,;CEJ;BOI;DI;:!/JH;D:
"7II>?<J;: (EI>?<J
"7II>?<J;: Correct inference Type 2 error:
(EI>?<J Type 1 error:
Correct inference
J?EDI E< ?D:?L?:K7B FB7DJ IF;9?;I <EH L7B?:7J?D= J>; &-* ;= 3K ;J 7B
 1>?J; ;J 7B  4>7D= ;J 7B  )KH I?CKB7J?ED H;IKBJI
KH=; 97KJ?ED ED IK9> 7FFHE79>;I J>; L7B?:7J?ED I>EKB: 8; 9ED:K9J;:
?D U;B:I M?J> 7 I?D=B; IF;9?;I EH <;M IF;9?;I M?J> 7 IJ78B; 9ECFEI?
J?ED .>?H: M; <EKD: J>7J ?J ?I B?A;BO J>7J 8EJ> J>; H;B7J?L; J>H;I>EB:
C;J>E: 7D: J>; 9KHL7JKH; C;J>E: M?BB ;HHED;EKIBO :;J;9J J>; 9>7D=;
E< J>; F?N;B !/ ;L;D ?< J>; ;D:C;C8;H !/I :E DEJ I>?<J .>?I H;IKBJ
IK==;IJI J>7J J>; !/ ?:;DJ?U97J?ED 87I;: EDBO ED J>; J>H;I>EB: EH J>;
:;H?L7J?L; E< J>; C?N;: 0# J;CFEH7B FHEUB; C7O DEJ 7:;GK7J;BO 97FJKH;
J>; EDI;J E< =H;;DD;II 7I J>;H; ?I 7 9EDI?:;H78B; :;=H;; E< KD9;HJ7?DJO
M?J>?D J>; B7D: IKH<79; 9ED:?J?EDI 7I M;BB 7I J>; 0# FHE:K9JI 8;?D= KI;:
";BC7D 
E< H;CEJ;BO I;DI;: !/ JE J>; C?N;: F?N;B ;<<;9J 1>?B; EL;H9EC?D= J>;
B78EH ?DJ;DIJ?L; D7JKH; E< U;B: MEHA J>; I?CKB7J?ED 7FFHE79> 97D 9ED
JHEB F>;DEBE=?97B L7H?78B;I M?J>?D 7 :;I?H78B; H7D=; .E EKH ADEMB;:=;
?J ?I H;B7J?L;BO 9>7BB;D=?D= JE ?:;DJ?<O =HEKD:87I;: !/I E< 9E;N?IJ?D=
IF;9?;I EL;H CKBJ?FB; O;7HI 7D: 9EDJHEBB?D= <EH J>;?H DED!/ L7H?78B;I
IK9> 7I J>; IF;9?;I 9ECFEI?J?ED J>; IE?B 879A=HEKD: J>; C7N?CKC
C?D?CKC (0# 7D: J>; =HEMJ> F;H?E: ?I D;7HBO ?CFEII?8B; 1>?B; J>;
97KJ?EDI 7H; D;;:;: ?HIJ J>; F7F;H ;N7C?D;I EDBO JME ;D:C;C8;HI 7I
?J ?I J>; I?CFB?;IJ 97I; E< 9E;NI?J?D= IF;9?;I ?D 7D ;9EIOIJ;C .E <79?B?
J7J; <KHJ>;H ;NFBEH7J?ED E< J>; FHE8B;C M; >7L; :;I?=D;: 7 I?CKB7J?ED
JEEBA?J JE ?D9BK:; 97I;I E< J>H;; EH CEH; ;D:C;C8;HI 7I I>EMD ?D J>;
-KFFB;C;DJ7HO 7J7 -;9ED: ?J I>EKB: 8; DEJ;: J>7J J>; 9ED9BKI?EDI
U;B: ;NF;H?C;DJI *EII?8B; C;J>E:I JE L7B?:7J; EKH UD:?D=I ?D9BK:; :;
J;9J?D= F>;DEBE=?97B 9>7D=;I M?J> <;M C?N;: IF;9?;I 7J 7 IC7BB IF7J?7B
I97B; 7D: 9ED:K9J?D= 9B?C7J;9EDJHEBB;: ;NF;H?C;DJI 1EBAEL?9> ;J 7B
 #D J>?I H;IF;9J J>; H;IKBJI :;H?L;: <HEC EKH 7D7BOJ?9 <H7C;MEHA
7D: J>; I?CKB7J?ED JEEBA?J M?BB <79?B?J7J; J>; :;I?=D E< IK9> U;B: ;NF;H
?C;DJI .>?H: J>;H; 7H; EJ>;H IEKH9;I E< C;7IKH;C;DJ KD9;HJ7?DJ?;I ?D
J>; FHE:K9J?ED E< 0# FHE:K9JI IK9> 7I J>; J?C; 9ECFEI?J?D= J>; E8I;H
L7J?ED =;EC;JHO 7D: J>; =H?::?D= 7HJ?<79JI -K9> KD9;HJ7?DJ?;I I>EKB:
8; 7D7BOP;: ?D <KJKH; 7JJ;CFJI
X. Chen et al. Remote Sensing of Environment xxx (2018) xxx-xxx
.>?I IJK:O ?I IKFFEHJ;: 8O J>; KD: <EH H;7J?L; ,;I;7H9> !HEKFI
E< (7J?ED7B (7JKH7B -9?;D9; EKD:7J?ED E< >?D7 (E  J>;
%;O ,;I;7H9> *HE=H7C E< HEDJ?;H -9?;D9;I !H7DJ (E
+34--1+ E< J>; >?D;I; 97:;CO E< -9?;D9;I 7 =H7DJ <HEC
J>; (7J?ED7B (7JKH7B -9?;D9; EKD:7J?ED E< >?D7 (E  7D:
J>; 3EKJ> #DDEL7J?ED *HECEJ?ED IIE9?7J?ED E< J>; >?D;I; 97:;CO E<
-9?;D9;I (E  .>; 7KJ>EHI 7FFH;9?7J; &?9ED= &?K <EH 7II?IJ?D=
M?J> :;I?=D?D= J>; I?CKB7J?ED JEEBA?J
Appendix A. Supplementary data
-KFFB;C;DJ7HO :7J7 JE J>?I 7HJ?9B; 97D 8; <EKD: EDB?D; 7J >JJFI
>B  !EM;H -. KHHEMI -( ->787DEL (0 'OD;D? , %DO7P?A>?D 3 
'ED?JEH?D= IFH?D= 97DEFO F>;DEBE=O E< 7 :;9?:KEKI 8HE7:B;7< <EH;IJ KI?D= ')#-
,;CEJ; -;DI DL?HED   T
C7D  ,7D:H?7C7D7DJ;D7 "* *E:7?H;  HEK?D ,  /FI97B; ?DJ;=H7J?ED E<
DEHC7B?P;: :?<<;H;D9; L;=;J7J?ED ?D:;N J>; FHE8B;C E< IF7J?7B >;J;HE=;D;?JO #
.H7DI !;EI9? ,;CEJ; -;DI   T
:7CI $ -C?J> ') $E>DIED *  -F;9JH7B C?NJKH; CE:;B?D= 7 D;M 7D7BOI?I
E< HE9A 7D: IE?B JOF;I 7J J>; 0?A?D= &7D:;H  I?J; $EKHD7B E< !;EF>OI?97B ,;I;7H9>
-EB?: 7HJ>   T
7:;9A 1 ED:;7K  öJJ9>;H % EAJEH  &K9>J 1 -9>78;H $ -?J9> - 
,;IFEDI;I E< IFH?D= F>;DEBE=O JE 9B?C7J; 9>7D=; (;M *>OJEB   T
HEMD ' ; ;KHI %' '7HI>7BB '  !BE87B F>;DEBE=?97B H;IFEDI; JE 9B?C7J;
9>7D=; ?D 9HEF 7H;7I KI?D= I7J;BB?J; H;CEJ; I;DI?D= E< L;=;J7J?ED >KC?:?JO 7D: J;C
F;H7JKH; EL;H  O;7HI ,;CEJ; -;DI DL?HED  T
K9>;C?D  !EK8?;H $ EKHH?;H !  'ED?JEH?D= F>;DEBE=?97B A;O IJ7=;I 7D: 9O
9B; :KH7J?ED E< J;CF;H7J; :;9?:KEKI <EH;IJ ;9EIOIJ;CI M?J> ()0",, :7J7 ,;
CEJ; -;DI DL?HED   T
!7E 2 "K;J; , (? 1 '?KH7 .  )FJ?97B8?EF>OI?97B H;B7J?EDI>?FI E< L;=
;J7J?ED IF;9JH7 M?J>EKJ 879A=HEKD: 9EDJ7C?D7J?ED ,;CEJ; -;DI DL?HED  
";BC7D   &7D: IKH<79; F>;DEBE=O 1>7J :E M; H;7BBO I;;<HEC IF79; .>; -9?
;D9; E< J>; .EJ7B DL?HEDC;DJ
"K;J;  ?:7D % '?KH7 . ,E:H?=K;P * !7E 2 ;HH;?H7 &!  )L;HL?;M
E< J>; H7:?EC;JH?9 7D: 8?EF>OI?97B F;H<EHC7D9; E< J>; ')#- L;=;J7J?ED ?D:?9;I ,;
CEJ; -;DI DL?HED   T
$öDIIED * ABKD:> &  .#'-.7 FHE=H7C <EH 7D7BOP?D= J?C;I;H?;I E< I7J;BB?J;
I;DIEH :7J7 ECFKJ !;EI9?   T
%;H:?B;I " !HED:ED7 ')  ()0",, (0# :;9ECFEI?J?ED 7D: IK8F?N;B
9B7II?U97J?ED KI?D= B?D;7H C?N?D= ?D J>; H=;DJ?D;7D *7CF7 #DJ $ ,;CEJ; -;DI 
 T
&? 4 EN $'  '7FF?D= HK88;H JH;; =HEMJ> ?D C7?DB7D: -EKJ>;7IJ I?7 KI?D=
J?C;I;H?;I ')#-  C (0# 7D: IJ7J?IJ?97B :7J7 FFB !;E=H   T
&?7D= & -9>M7HJP ' ;? -&  07B?:7J?D= I7J;BB?J; F>;DEBE=O J>HEK=> ?DJ;D
I?L; =HEKD: E8I;HL7J?ED 7D: B7D:I97F; I97B?D= ?D 7 C?N;: I;7IED7B <EH;IJ ,;CEJ;
-;DI DL?HED  T
'7JIKCEJE % )>J7 . #H7I7M7 ' (7A7CKH7 .  B?C7J; 9>7D=; 7D: ;NJ;D
I?ED E< J>; !?DA=E 8?BE87 & =HEM?D= I;7IED ?D $7F7D !BE8 >7D= ?EB  
'EDJ7D:ED &' -C7BB   .>; ?CF79J E< IE?B H;V;9J7D9; ED J>; GK7DJ?U97J?ED E<
J>; =H;;D L;=;J7J?ED <H79J?ED <HEC (0# ,;CEJ; -;DI DL?HED   T
,7JAEMIAO   (EDB?D;7H ,;=H;II?ED 'E:;B?D= /D?U;: *H79J?97B FFHE79>
'7H9;B ;AA;H (;M 3EHA T
,;;:  HEMD $  07D:;H4;;  &EL;B7D: ., ';H9>7DJ $1 )>B;D ) 
';7IKH?D= F>;DEBE=?97B L7H?78?B?JO <HEC I7J;BB?J; ?C7=;HO $ 0;= -9?   T
-9>M7HJP ' ,;;:   -KH<79; F>;DEBE=O 7D: I7J;BB?J; I;DIEH:;H?L;: EDI;J E<
=H;;DD;II 7D ?D?J?7B 9ECF7H?IED #DJ $ ,;CEJ; -;DI   T
->;D ' .7D= 3 >;D $ 4>K 2 4>;D= 3  #DVK;D9;I E< J;CF;H7JKH; 7D: FH;
9?F?J7J?ED 8;<EH; J>; =HEM?D= I;7IED ED IFH?D= F>;DEBE=O ?D =H7IIB7D:I E< J>; 9;DJH7B
7D: ;7IJ;HD +?D=>7?.?8;J7D *B7J;7K =H?9 EH ';J;EHEB  T
->;D ' -KD 4 17D= - 4>7D= ! %ED= 1 >;D  *?7E -  (E ;L?:;D9;
E< 9EDJ?DKEKIBO 7:L7D9;: =H;;DKF :7J;I ?D J>; .?8;J7D *B7J;7K EL;H J>; B7IJ :;97:;
*HE9 (7JB 97: -9?   
->;D ' .7D= 3 >;D $ 37D= 2 17D=  K? 2 37D= 3 "7D & &? & K $
4>7D= ! ED= (  7HB?;HI;7IED L;=;J7J?ED >7I =H;7J;H J;CF;H7JKH; I;DI?J?L
?JO E< IFH?D= F>;DEBE=O ?D DEHJ>;HD >;C?IF>;H; *&E- )D;  ;
-Jö9AB? , 0?:7B; *&  KHEF;7D FB7DJ F>;DEBE=O 7D: 9B?C7J; 7I I;;D ?D 7 O;7H
0",, B7D:IKH<79; F7H7C;J;H :7J7I;J #DJ $ ,;CEJ; -;DI   T
-OA;I '. *H;DJ?9; #  B?C7J; 9>7D=; JH;; IF;9?;I :?IJH?8KJ?EDI 7D: <EH;IJ :O
D7C?9I 7 97I; IJK:O ?D J>; C?N;: 9ED?<;HDEHJ>;HD >7H:MEE:I PED; E< DEHJ>;HD K
HEF; B?C >7D=   T
.H;C8B7O () &7HE9GK; !,  -;7IED7B :OD7C?9I E< KD:;HIJEHO L;=;J7J?ED ?D <EKH
;7IJ;HD 7D7:?7D <EH;IJ JOF;I #DJ $ *B7DJ -9?   T
1>?J; ' .>EHDJED * ,KDD?D= -1   9EDJ?D;DJ7B F>;DEBE=O CE:;B <EH
CED?JEH?D= L;=;J7J?ED H;IFEDI;I JE ?DJ;H7DDK7B 9B?C7J?9 L7H?78?B?JO !BE8 ?E
=;E9>;C O9B;I   T
1>?J; ' ;KHI  %?HIJ;D ' ?:7D % #DEKO; 1 ,?9>7H:IED  S &;;KM;D
0  #DJ;H9ECF7H?IED ?DJ;HFH;J7J?ED 7D: 7II;IIC;DJ E< IFH?D= F>;DEBE=O ?D
(EHJ> C;H?97 ;IJ?C7J;: <HEC H;CEJ; I;DI?D= <EH  !BE8 >7D= ?EB 
 T
1EBAEL?9> ' EEA # BB;D $' H?CC?DI .' ;J7D9EKHJ $& .H7L;HI -
*7K - ,;=;JP $ 7L?;I .$ %H7<J ($ KBJ ., EBC=H;D % '7P;H -$ '9
78; !$ '9!?BB $ *7HC;I7D  -7B7C?D ( -9>M7HJP ' B;B7D: 
 17HC?D= ;NF;H?C;DJI KD:;H FH;:?9J FB7DJ F>;DEBE=?97B H;IFEDI;I JE 9B?C7J;
9>7D=; (7JKH;  T
3K " &K;:;B?D=  2K $  1?DJ;H 7D: IFH?D= M7HC?D= H;IKBJ ?D :;B7O;: IFH?D=
F>;DEBE=O ED J>; .?8;J7D *B7J;7K *HE9 (7JB 97: -9?   T
4>7D= 2 !EB:8;H= '  'ED?JEH?D= <7BB <EB?7=; 9EBEH7J?ED :OD7C?9I KI?D=
J?C;I;H?;I I7J;BB?J; :7J7 ,;CEJ; -;DI DL?HED   T
4>7D= 2 H?;:B ' -9>77<  -JH7>B;H " "E:=;I $ !7E  S "K;J; 
 'ED?JEH?D= L;=;J7J?ED F>;DEBE=O KI?D= ')#- ,;CEJ; -;DI DL?HED  
4>7D= 2 H?;:B ' -9>77<   !BE87B L;=;J7J?ED F>;DEBE=O <HEC 'E:;H
7J; ,;IEBKJ?ED #C7=?D= -F;9JHEH7:?EC;J;H ')#- ;L7BK7J?ED E< =BE87B F7JJ;HDI 7D:
9ECF7H?IED M?J> ?D I?JK C;7IKH;C;DJI $EKHD7B E< !;EF>OI?97B ,;I;7H9> ?E=;E
I9?;D9;I  !
4>7D= ! 4>7D= 3 ED= $ 2?7E 2  !H;;DKF :7J;I ?D J>; .?8;J7D *B7J;7K
>7L; 9EDJ?DKEKIBO 7:L7D9;: <HEC  JE  *HE9 (7JB 97: -9?  
4>7D= 2 17D= $ !7E  &?K 3 -9>77<  H?;:B ' S 37D   NFBEH7J?ED
E< I97B?D= ;<<;9JI ED 9E7HI; H;IEBKJ?ED B7D: IKH<79; F>;DEBE=O ,;CEJ; -;DI DL?HED
 T
4>7E . -9>M7HJP '  N7C?D?D= J>; EDI;J E< IFH?D= ?D 1?I9EDI?D B?C ,;I 
 T
... Therefore, earlier leaf unfolding would also lead to the advance of programmed leaf life cycle and further result in different changes of V NDVI in early and late GUP. As our analyses were conducted based on grid NDVI product, the mixed pixel effect would make the changes of V NDVI more complicated (Chen et al., 2018). For instance, it is well known that increasing temperature can accelerate plant growth via temperature-dependent enzymatic catalytic reactions if temperature is lower than temperature optima (Atkinson & Porter, 1996). ...
... However, the optimal temperature of different species in a same pixel may be various. The different temperature optima and asynchronous movements of growth cycles among different species would generate a nonlinear response of V NDVI to climate change at landscape or larger levels (Chen et al., 2018). ...
Full-text available
Global vegetation greening has been widely confirmed in previous studies, yet the changes in the velocity of green‐up in each month of green‐up period (GUP) remains unclear. Here, we defined the velocity of vegetation green‐up as VNDVI (the monthly increase of Normalized Difference Vegetation Index (NDVI) during GUP) and further explored its response to climate change in middle‐high latitude Northern Hemisphere. We found that in early GUP, VNDVI generally showed positive trends from 1982 to 2015, while in late GUP, it showed negative trends in most areas. Such contrasting trends were mainly due to a positive temperature effect on VNDVI in early GUP, but this effect turned negative in late GUP. The increase of soil moisture also in part explained the accelerated vegetation green‐up, especially in the arid and semi‐arid ecosystems of inland areas. Our analyses also indicate that the first month of the GUP was the key stage impacting vegetation greenness in summer. Future warming may continuously speed up the early growth of vegetation, altering the seasonal trajectory of vegetation and its feedbacks to the Earth system.
... Remote sensing (RS) has been shown to have great potential and advantages in earth surface monitoring for several decades [1][2][3]. Most of the current satellites have low to moderate spatial resolution, meaning that one pixel usually contains a mixture of vegetation (tree and understory) and background (soil, shade, etc.), which is known as the mixed-pixel problem [4][5][6]. This issue is particularly significant for semi-arid and dryland ecosystems where the vegetation is usually sparse [7]. ...
... Approaches for solving this problem include reducing soil signals at low vegetation cover by adding soil brightness correction factors (SAVI, MSAVI2, OSAVI, GSAVI) [24][25][26][27] or enhancing vegetation signals at high vegetation cover by adding weighting coefficients (EVI, WDRVI, NIRv) [20,21,28]. However, these indices that are derived from most satellites still may not accurately capture surface phenological changes due to their limited spatial resolutions [6]. In addition, shade also causes biased NDVI values, thus significantly reducing land cover classification accuracy. ...
Full-text available
Remote sensing (RS) for vegetation monitoring can involve mixed pixels with contributions from vegetation and background surfaces, causing biases in signals and their interpretations, especially in low-density forests. In a case study in the semi-arid Yatir forest in Israel, we observed a mismatch between satellite (Landsat 8 surface product) and tower-based (Skye sensor) multispectral data and contrasting seasonal cycles in near-infrared (NIR) reflectance. We tested the hypothesis that this mismatch was due to the different fractional contributions of the various surface components and their unique reflectance. Employing an unmanned aerial vehicle (UAV), we obtained high-resolution multispectral images over selected forest plots and estimated the fraction, reflectance, and seasonal cycle of the three main surface components (canopy, shade, and sunlit soil). We determined that the Landsat 8 data were dominated by soil signals (70%), while the tower-based data were dominated by canopy signals (95%). We then developed a procedure to resolve the canopy (i.e., tree foliage) normalized difference vegetation index (NDVI) from the mixed satellite data. The retrieved and corrected canopy-only data resolved the original mismatch and indicated that the spatial variations in Landsat 8 NDVI were due to differences in stand density, while the canopy-only NDVI was spatially uniform, providing confidence in the local flux tower measurements.
... For decades, a series of classifiers has been developed to interpret remote sensing images by allocating each pixel to a single-class label . However, the signal (i.e., spectrum composed of the reflectance of all spectral bands) of a pixel is generally composed of spectra from several land cover types, known as the "mixed-pixel problem" (Chen et al., 2018a). Moreover, the inherent point spread function effect of sensors can aggravate the mixed pixel problem (Wang et al., 2020b). ...
Full-text available
Spatio-temporal subpixel mapping (STSPM) has shown great potential for monitoring land surfaces, by generating land cover maps with both fine spatial and temporal resolutions. Selecting cloud-free fine spatial resolution images as ancillary data for STSPM can ensure that the temporal dependence term is measured for all subpixels, as in all current STSPM methods. However, such images are generally limited by cloud contamination, thereby resulting in great land cover changes between the available clear image and the desired fine spatial resolution land cover map. This research proposes a cloud-independent STSPM (C-STSPM) method to reconstruct the fine spatial resolution land cover maps by using cloudy images directly, which are assumed to have fewer land cover changes than temporally distant clear images. Cloud-independent spatio-temporal dependence was proposed in the presence of cloudy pixels. Experiments were performed under various cloud conditions involving 21 × 21 pairs of simulated cloudy images. The results demonstrate that by utilizing land cover information of clear pixels in cloudy images, more accurate prediction can be produced by C-STSPM compared to directly discarding those cloudy images, even if the number of cloud pixels increases to 95%. The advantage of C-STSPM is more evident when the clouds are distributed sparsely, which benefits from the increased number of clear pixels at the edge of the cloudy areas. Furthermore, a negative linear correlation was detected between the prediction accuracy and the ratio of overlapping cloudy pixels in the cloudy images. Moreover, the C-STSPM method helps to deal with abrupt changes occurred in the temporally distant cloud-free images by utilizing the temporally adjacent cloudy images with gradual land cover changes. Overall, the C-STSPM method provides a completely new solution to make fuller use of the widely existing cloudy images in multi-scale time-series images.
... Wu et al. (2017) [19] compared different satellite-based methods for estimating the growing season onset, and all of them showed relatively low correlation with the observed results of eddy flux measured at 60 sites. Many reasons will lead to the deviation of prediction results [27,37], but the main issue, i.e., the scale-up vs. mixed-pixel problem, is hard to shift, but perhaps unavoidable [26,38]. Although the spatial resolution of satellite-derived vegetation index products has been improved in recent years, a given pixel may include a combination of different vegetation types that might result in significantly different phenology, which challenges the determination of a reliable onset date of spring phenological phases within a heterogeneous ecoregion. ...
Full-text available
Multiple methods have been developed to identify the transition threshold from the reconstructed satellite-derived normalized difference vegetation indices (NDVI) time series and to determine the inflection point corresponding to a certain phenology phase (e.g., the spring green-up date (GUD)). We address an issue that large uncertainties might occur in the inflection point identification of spring GUD using the traditional satellite-based methods since different vegetation types exhibit asynchronous phenological phases over a heterogeneous ecoregion. We tentatively developed a Maximum-derivative-based (MDB) method and provided inter-comparisons with two traditional methods to detect the turning points by the reconstructed time-series data of NDVI for identifying the GUD against long-term observations from the sites covered by a mixture of deciduous forest and herbages in the Pan European Phenology network. Results showed that higher annual mean temperature would advance the spring GUD, but the sensitive magnitudes differed depending on the vegetation type. Therefore, the asynchronization of phenological phases among different vegetation types would be more pronounced in the context of global warming. We found that the MDB method outperforms two other traditional methods (the 0.5-threshold-based method and the maximum-ratio-based method) in predicting the GUD of the subsequent-green-up vegetation type when compared with ground observation, especially at sites with observed GUD of herbages earlier than deciduous forest, while the Maximum-ratio-based method showed better performance for identifying GUDs of the foremost-green-up vegetation type. Although the new method improved in our study is not universally applicable on a global scale, our results, however, highlight the limitation of current inflection point identify algorithms in predicting the GUD derived from satellite-based vegetation indices datasets in an ecoregion with heterogeneous vegetation types and asynchronous phenological phases, which makes it helpful for us to better predict plant phenology on an ecoregion-scale under future ongoing climate warming.
... The use of UAV observations and tower-mounted cameras can, to a certain extent, remediate the scale mismatch issue [27,28]. Meanwhile, considering the complexity of scale effects, computer simulations based on 3D radiative-transfer modelling can be used as a powerful tool to explore the scale effects or mixed image effects in vegetation phenology remote-sensing monitoring [29]. In addition, for low-and medium-resolution remote-sensing phenology products (e.g., MODIS/VIIRS), it is difficult even for UAVs or phenocams to provide validation data at the comparable pixel scale, in which case indirect "validation" can be performed using higher-spatial-resolution satellite data [30]. ...
Full-text available
Land surface phenology (LSP) is an important research field in terrestrial remote sensing and has become an indispensable approach in global change research, as evidenced by many important scientific findings supported by LSP in recent decades. LSP involves the use of remote sensing to monitor seasonal dynamics in vegetated land surfaces and to retrieve phenological metrics (transition dates, rate of change, annual integrals, etc.). LSP is an essential indicator of global change and has played a pivotal role in shaping our understanding about how terrestrial ecosystems are responding to climate change and human activities. Both regional and global LSP products have been routinely generated and played prominent roles in modeling crop yield, ecological surveillance, identifying invasive species, modeling the terrestrial biospheric processes, and assessing global change impacts on urban and natural ecosystems. Recent advances in field and spaceborne sensor technologies, as well as data fusion techniques, have enabled novel LSP retrieval algorithms that refine LSP retrievals at even higher spatiotemporal resolutions, providing new insights into ecosystem dynamics. Meanwhile, rigorous assessment of the uncertainties in LSP retrievals is undergoing, and efforts to reduce these uncertainties are also forming an active research field. In addition, open-source software and hardware are being developed and have greatly facilitated the use of LSP metrics by scientists beyond the remote-sensing community. As such, we organized this Special Issue to cover the latest developments in sensor technologies, LSP retrieval algorithms and validation strategies, and the use of LSP products in a variety of fields. The objective of this Editorial is to offer the readers an overview of the latest developments in the LSP field and facilitate the distribution of the scientific knowledge from this Special Issue.
... For any type of image data, pixel mixing is a common phenomenon [31][32][33]. A mixed pixel may contain an unknown composition of land cover types [34], affecting the simulation of radiative characteristics and inversion for land surface parameters from remote sensing data [33]. Mixing pixels has been a barrier in the application of remote sensing to burned area mapping. ...
Full-text available
The accurate detection of burned forest area is essential for post-fire management and assessment, and for quantifying carbon budgets. Therefore, it is imperative to map burned areas accurately. Currently, there are few burned-area products around the world. Researchers have mapped burned areas directly at the pixel level that is usually a mixture of burned area and other land cover types. In order to improve the burned area mapping at subpixel level, we proposed a Burned Area Subpixel Mapping (BASM) workflow to map burned areas at the subpixel level. We then applied the workflow to Sentinel 2 data sets to obtain burned area mapping at subpixel level. In this study, the information of true fire scar was provided by the Department of Emergency Management of Hunan Province, China. To validate the accuracy of the BASM workflow for detecting burned areas at the subpixel level, we applied the workflow to the Sentinel 2 image data and then compared the detected burned area at subpixel level with in situ measurements at fifteen fire-scar reference sites located in Hunan Province, China. Results show the proposed method generated successfully burned area at the subpixel level. The methods, especially the BASM-Feature Extraction Rule Based (BASM-FERB) method, could minimize misclassification and effects due to noise more effectively compared with the BASM-Random Forest (BASM-RF), BASM-Backpropagation Neural Net (BASM-BPNN), BASM-Support Vector Machine (BASM-SVM), and BASM-notra methods. We conducted a comparison study among BASM-FERB, BASM-RF, BASM-BPNN, BASM-SVM, and BASM-notra using five accuracy evaluation indices, i.e., overall accuracy (OA), user’s accuracy (UA), producer’s accuracy (PA), intersection over union (IoU), and Kappa coefficient (Kappa). The detection accuracy of burned area at the subpixel level by BASM-FERB’s OA, UA, IoU, and Kappa is 98.11%, 81.72%, 74.32%, and 83.98%, respectively, better than BASM-RF’s, BASM-BPNN’s, BASM-SVM’s, and BASM-notra’s, even though BASM-RF’s and BASM-notra’s average PA is higher than BASM-FERB’s, with 89.97%, 91.36%, and 89.52%, respectively. We conclude that the newly proposed BASM workflow can map burned areas at the subpixel level, providing greater accuracy in regards to the burned area for post-forest fire management and assessment.
The spatiotemporal variations in spring onset were evaluated based on a satellite-derived enhanced vegetation index (EVI), land surface temperature, and snow cover from 2000 to 2021 in interior Alaska. Spring offset was determined as a budburst or seasonal green-up of vegetation based on EVI. The analysis showed that snow disappearance date was the key driver of spatiotemporal variations in the spring onset date. These dates for snow disappearance and spring onset varied with elevation, showing importance of elevation to explain the spatial heterogeneity in spring onset. The magnitudes in the heterogeneity in spring onset varied year to year, where the heterogeneity increased in warm spring owing to an increased heterogeneity in snow disappearance. This phenomenon was associated with slower snowmelt under lower shortwave radiation earlier in the season. These results suggest that future warming could increase the heterogeneity in spring onset in this region. Spring onset was further modeled with the growing degree-day model using the Bayesian optimization, which showed that model performance did not deteriorate with coarse-resolution inputs. The results indicate that coarse-resolution models, such as Earth system models, are a reliable tool for predicting spring onset in this region.
Vegetation phenology influences many ecosystem and climate processes, such as carbon uptake and energy and water cycles. Thus, understanding drivers of vegetation phenology is crucial for predicting current and future impacts of climate change on ecological systems. Existing models can accurately predict the date of spring green-up in temperate forests but tend to perform poorly in grassland systems. We hypothesize this is because most do not incorporate water availability, a primary limiting factor for grassland plants. In this study, we used long-term datasets of digital imagery from the PhenoCam Network of 43 diverse North American grassland sites (195 site-years) to test existing spring phenology models, as well as develop several new models that incorporate precipitation or soil moisture (53 models). Most of the new models performed substantially better, with the best model requiring sufficient accumulated precipitation followed by warm temperatures to trigger spring onset (root mean square error, RMSE, between predicted and observed dates = 16.0 days). Importantly, the best model performed well across all grassland types using a single set of parameters, from temperate to arid grasslands. Since plants are adapted to their local climates, model performance was further improved when parameters were independently optimized for four separate climate regions (RMSE = 10.4 days). Therefore, both sufficient precipitation and temperature are required for grassland green-up, but optimal thresholds vary by region. Running the top model with projected climate data (representative concentration pathway 8.5) suggests that, depending on the climate region, spring onset will occur up to 12 days earlier within 100 years in temperature-limited sites, but the trend is unclear for precipitation-limited sites (3.5 ± 8.0 days later). This new phenology model improves our ability to understand and predict grassland dynamics, with implications for both current and future ecosystem processes related to carbon and water cycling.
Full-text available
Vegetation phenology has been viewed as the nature’s calendar and an integrative indicator of plant-climate interactions. The correct representation of vegetation phenology is important for models to accurately simulate the exchange of carbon, water, and energy between the vegetated land surface and the atmosphere. Remote sensing has advanced the monitoring of vegetation phenology by providing spatially and temporally continuous data that in together with conventional ground observations offering a unique contribution to our knowledge about environmental impact on ecosystems as well as the ecological adaptations and feedback to global climate change. Land surface phenology is defined as the use of satellites to monitor seasonal dynamics in vegetated land surfaces and to estimate phenological transition dates. Land surface phenology, as an interdisciplinary subject among remote sensing, ecology and biometeorology, has undergone rapid development over the past few decades. Recent advances in sensor technologies, as well as data fusion techniques, have enabled novel phenology retrieval algorithms that refine phenology details at even higher spatiotemporal resolutions, providing new insights into ecosystem dynamics. As such, here we summarize the recent advances in land surface phenology and the associated opportunities for science applications. We focus on the remaining challenges, promising techniques, and emerging topics that together we believe will truly form the very frontier of global land surface phenology research field.
Greenhouse gases (GHGs) are major contributors to global warming and climate change. These gases modulate the atmospheric radiative forcing and play an important role in Earth's albedo. The emission level, global warming potential and the persistence of a GHG define its accumulation in the atmosphere and relative potential to change radiative forcing. The major anthropogenic GHGs include methane, nitric oxide, ozone, hydrochloroflourocarbons, chloroflourocarbons, sulfur hexaflouride and nitrogen triflouride. Besides these, some gases indirectly act as GHGs like carbon monoxide, non-methane hydrocarbons, and nitrogen oxides. Many scientists have already warned regarding elevated emission trends after the industrial revolution. From last decades the emission of GHGs has tremendously increased in the atmosphere and the natural sinks of GHGs have contracted over time. Generally, fossil fuel burning and change in land use are major sources of GHGs while major sinks include soil, ocean and atmosphere. Interestingly the emission trends of greenhouse gases from different sources as well as the contribution of various countries to global greenhouse gasses budget have changed. Thus previous footprints, trends and projections regarding GHGs are needed to be reevaluated. Specific precautions and strategies are compatible to reduce GHGs emissions while further may help to obtain global temperature to above pre-industrial ambient temperature level by reducing 2°C in current temperature.
Full-text available
In recent decades, satellite-derived start of vegetation growing season (SOS) has advanced in many northern temperate and boreal regions. Both the magnitude of temperature increase and the sensitivity of the greenness phenology to temperature-the phenological change per unit temperature-can contribute the advancement. To determine the temperature-sensitivity, we examined the satellite-derived SOS and the potentially effective pre-season temperature (T eff) from 1982 to 2008 for vegetated land between 30°N and 80°N. Earlier season vegetation types, i.e., the vegetation types with earlier SOSmean (mean SOS for 1982-2008), showed greater advancement of SOS during 1982-2008. The advancing rate of SOS against year was also greater in the vegetation with earlier SOSmean even the T eff increase was the same. These results suggest that the spring phenology of vegetation may have high temperature sensitivity in a warmer area. Therefore it is important to consider temperature-sensitivity in assessing broad-scale phenological responses to climatic warming. Further studies are needed to explore the mechanisms and ecological consequences of the temperature-sensitivity of start of growing season in a warming climate.
Full-text available
Expanding global and regional markets are driving the conversion of traditional subsistence agricultural and occupied non-agricultural lands to commercial-agricultural purposes. In many parts of mainland Southeast Asia rubber plantations are expanding rapidly into areas where the crop was not historically found. Over the last several decades more than one million hectares of land have been converted to rubber trees in areas of China, Laos, Thailand, Vietnam, Cambodia and Myanmar, where rubber trees were not traditionally grown. This expansion of rubber plantations has replaced ecologically important secondary forests and traditionally managed swidden fields and influenced local energy, water and carbon fluxes. Accurate and up-to-date monitoring and mapping of rubber tree growth is critical to understanding the implications of this changing ecosystem. Discriminating rubber trees from second-growth forests and fallow land has proven challenging. Previous experiments using machine-learning approaches with hard classifications on remotely sensed data, when faced with the realities of a heterogeneous plant-life mixture and high intra-class variance, have tended to overestimate the areas of rubber tree growth. Our current research sought to: 1) to investigate the potential of using a Mahalanobis typicality model to deal with mixed pixels; and 2) to explore the potential for combining MOderate Resolution Imaging Spectroradiometer (MODIS) imagery with sub-national statistical data on rubber tree areas to map the distribution of rubber tree growth across this mainland Southeast Asia landscape. Our study used time-series MODIS Terra 16-day composite 250 m Normalized Difference Vegetation Index (NDVI) products (MOD13Q1) acquired between March 2009 and May 2010. We used the Mahalanobis typicality method to identify pixels where rubber tree growth had the highest probability of occurring and sub-national statistical data on rubber tree growth to quantify the number of pixels of rubber tree growth mapped per administrative unit. We used Relative Operating Characteristic (ROC) and error matrix analysis, respectively, to assess the viability of Mahalanobis typicalities and to validate classification accuracy. High ROC values, over 0.8, were achieved with the Mahalanobis typicality images of both mature and young rubber trees. The proposed method greatly reduced the commission errors for the two types of rubber tree growth to 1.9% and 2.8%, respectively (corresponding to user’s accuracies of 98.1% and 97.2%, respectively). Results indicate that integrating Mahalanobis typicalities with MODIS time-series NDVI data and sub-national statistics can successfully overcome the earlier overestimation problem.
Full-text available
In the last two decades the availability of global remote sensing data sets has provided a new means of studying global patterns and dynamics in vegetation. The vast majority of previous work in this domain has used data from the Advanced Very High Resolution Radiometer, which until recently was the primary source of global land remote sensing data. In recent years, however, a number of new remote sensing data sources have become available that have significantly improved the capability of remote sensing to monitor global ecosystem dynamics. In this paper, we describe recent results using data from NASA's Moderate Resolution Imaging Spectroradiometer to study global vegetation phenology. Using a novel new method based on fitting piecewise logistic models to time series data from MODIS, key transition dates in the annual cycle(s) of vegetation growth can be estimated in an ecologically realistic fashion. Using this method we have produced global maps of seven phenological metrics at 1-km spatial resolution for all ecosystems exhibiting identifiable annual phenologies. These metrics include the date of year for (1) the onset of greenness increase (greenup), (2) the onset of greenness maximum (maturity), (3) the onset of greenness decrease (senescence), and (4) the onset of greenness minimum (dormancy). The three remaining metrics are the growing season minimum, maximum, and summation of the enhanced vegetation index derived from MODIS. Comparison of vegetation phenology retrieved from MODIS with in situ measurements shows that these metrics provide realistic estimates of the four transition dates identified above. More generally, the spatial distribution of phenological metrics estimated from MODIS data is qualitatively realistic, and exhibits strong correspondence with temperature patterns in mid- and high-latitude climates, with rainfall seasonality in seasonally dry climates, and with cropping patterns in agricultural areas.
Land surface phenology (LSP) provides bio-indication of ongoing climate change. It uses space-borne greenness proxies to monitor plant phenology at the landscape level from the regional to global scale. However, several un- considered methodological and observational -related limitations may lead to misinterpretation of the satellite- derived signals. For instance, changes in species composition within a pixel could result in a change in the time series of the greenness proxy, due to the distinct phenology of the plant species involved. The change in the signal would then be misinterpreted as a phenological change while it is actually related to changes in species composition within the pixel. Other limitations include the selection of the smoothing technique and the method used to extract the LSP metrics. These not only may affect the timing of the LSP metrics but also the sign of the observed LSP change. Another and much less known limitation is related to the mixed signal from multi-canopy layers. Satellites may detect changes that corresponds to the understorey layer in complex vertical vegetation systems while the ‘real’ contribution of this layer in terms of ecosystem functioning and dynamics might be small compared to the undetected overstorey layer in cases of a late overstorey development. Here, some of the LSP basics are reviewed with emphasis on these and other potential sources of misinterpretation. Several aids to overcome these limitations, which include suggestions for multi methods analysis and the integration of information from satellite and ground-based sensors are provided alongside some prospective future LSP research directions.
Numerous land surface phenology (LSP) datasets have been produced from various coarse resolution satellite data and different detection algorithms from regional to global scales. In contrast to field-observed phenological events that are defined by clearly evident organismal changes with biophysical meaning, current approaches to detecting transitions in LSP only determine the timing of variations in remotely sensed observations of surface greenness. Since activities to bridge LSP and field observations are challenging and limited, our understanding of the biophysical characteristics of LSP transitions is poor. Therefore, we set out to explore the scaling effects on LSP transitions at the nominal start of growing season (SOS) by comparing detections from coarse resolution data with those from finer resolution imagery. Specifically, using a hybrid piecewise-logistic-model-based LSP detection algorithm, we detected SOS in the agricultural core of the United States—central Iowa—at two scales: first, at a finer scale (30 m) using reflectance generated by fusing MODIS data with Landsat 8 OLI data (OLI SOS) and, second, at a coarser resolution of 500 m using Visible Infrared Imaging Radiometer Suite (VIIRS) observations. The VIIRS SOS data were compared with OLI SOS that had been aggregated using a percentile approach at various degrees of heterogeneity. The results revealed the complexities of SOS detections and the scaling effects that are latent at the coarser resolution. Specifically, OLI SOS variation defined using standard deviation (SD) was as large as 40 days within a highly spatially heterogeneous VIIRS pixel; whereas, SD could be < 10 days for a more homogeneous set of pixels. Furthermore, the VIIRS SOS detections equaled the OLI SOS (with an absolute difference less than one day) in > 60% of OLI pixels within a homogeneous VIIRS pixel, but in < 20% of OLI pixels within a spatially heterogeneous VIIRS pixel. Moreover, the SOS detections in a coarser resolution pixel reflected the timing at which vegetation greenup onset occurred in 30% of area, despite variation in SOS heterogeneities. This result suggests that (1) the SOS detections at coarser resolution are controlled more by the earlier SOS pixels at the finer resolution rather than by the later SOS pixels, and (2) it should be possible to well simulate the coarser SOS value by selecting the timing at 30th percentile SOS at the finer resolution. Finally, it was demonstrated that in homogeneous areas the VIIRS SOS was comparable with OLI SOS with an overall difference of < 5 days.
Regional phenology is important in ecosystem simulation models and coupled biosphere/atmosphere models. In the continental United States, the timing of the onset of greenness in the spring (leaf expansion, grass green-up) and offset of greenness in the fall (leaf abscission, cessation of height growth, grass brown-off) are strongly influenced by meteorological and climatological conditions. We developed predictive phenology models based on traditional phenology research using commonly available meteorological and climatological data. Predictions were compared with satellite phenology observations at numerous 20km×20km contiguous landcover sites. Onset mean absolute error was 7.2 days in the deciduous broadleaf forest (DBF) biome and 6.1 days in the grassland biome. Offset mean absolute error was 5.3 days in the DBF biome and 6.3 days in the grassland biome. Maximum expected errors at a 95% probability level ranged from 10 to 14 days. Onset was strongly associated with temperature summations in both grassland and DBF biomes; DBF offset was best predicted with a photoperiod function, while grassland offset required a combination of precipitation and temperature controls. A long-term regional test of the DBF onset model captured field-measured interannual variability trends in lilac phenology. Continental application of the phenology models for 1990-1992 revealed extensive interannual variability in onset and offset. Median continental growing season length ranged from a low of 129 days in 1991 to a high of 146 days in 1992. Potential uses of the models include regulation of the timing and length of the growing season in large-scale biogeochemical models and monitoring vegetation response to interannual climatic variability.
Fall foliage coloration is a phenomenon that occurs in many deciduous trees and shrubs worldwide. Measuring the phenology of fall foliage development is of great interest for climate change, the carbon cycle, ecology, and the tourist industry; but little effort has been devoted to monitoring the regional fall foliage status using remotely-sensed data. This study developed an innovative approach to monitoring fall foliage status by means of temporally-normalized brownness derived from MODIS (Moderate Resolution Imaging Spectroradiometer) data. Specifically, the time series of the MODIS Normalized Difference Vegetation Index (NDVI) was smoothed and functionalized using a sigmoidal model to depict the continuous dynamics of vegetation growth. The modeled temporal NDVI trajectory during the senescent phase was further combined with the mixture modeling to deduce the temporally-normalized brownness index which was independent of the surface background, vegetation abundance, and species composition. This brownness index was quantitatively linked with the fraction of colored and fallen leaves in order to model the fall foliage coloration status. This algorithm was tested by monitoring the fall foliage coloration phase using MODIS data in northeastern North America from 2001 to 2004. The MODIS-derived timing of foliage coloration phases was compared with in-situ measurements, which showed an overall absolute mean difference of less than 5days for all foliage coloration phases and about 3days for near peak coloration and peak coloration. This suggested that the fall foliage coloration phase retrieved from the temporally-normalized brownness index was qualitatively realistic and repeatable.
Understory vegetation dynamics was monitored throughout the 1998 growing season in four eastern Canadian forest types: sugar maple-American beech (Acer saccharum Marsh.-Fagus grandifolia Ehrh.); sugar maple-yellow birch (Betula alleghaniensis Britton); balsam fir (Abies balsamea [L.] Mill.); and black spruce-jack pine (Picea mariana [Mill.] BSP-Pinus banksiana Lamb.). Significant differences in biomass were obtained among species groups (herbaceous, woody, and mosses) within each site. However, biomass did not vary significantly throughout the growing season, except in the sugar maple-American beech site. The four sites differed in total biomass, but these differences could be explained mainly by the presence of some key species (e.g., Lycopodium lucidulum Michx.) and the light regime under the canopy. Species richness varied throughout the growing season on each site. However, peaks in richness did not occur at the same time in each site: richness peaked later with increasing latitude, from late May in the southernmost site to late September in the northernmost site. Average richness did not vary across sites. Aboveground nutrient concentrations for herbaceous species were greater than nutrient concentrations for stems of woody species. However, nutrient concentrations for woody species leaves were comparable to those of herbaceous species. Our results highlighted the importance of understory species in the cycling of nutrients and their capacity to keep nutrients within a site. Specific leaf area (SLA) for herbaceous species was greater than SLA for woody species, which indicated a greater capacity to acclimate to different light conditions than woody species. Increase in leaf area ratio with decrease in height for sugar maple and beech suggests that both species decreased their efficiency to produce biomass at the ground level as they were overtopped by higher surrounding vegetation.