Monitoring Agricultural Drought in South Korea - Monitoring and Predicting Agricultural Drought: A Global Study - page 394

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46

orbiting satellites to monitor vegetation conditions in Korea in 1999 (figure
30.6). These maps were used to classify the land cover of rice at a 1-km
grid spacing for land surface parameterization of the biosphere model (Koo
et al., 2001). Ha et al. (2001) analyzed the temporal variability in the
NDVI, leaf area index (LAI), and surface temperature ( T s) estimated from
AVHRR data collected from Korean Peninsula during 1981-1994. These
products can be applied to estimate AET (Szilagyi, 2002):

¯ E gs

AET est = NDVI (σ gs ) + [30.1]

where AET est is the estimated AET, σ gs is the standard deviation of monthly
AET during the growing season (mm/day), and E gs is mean growing season
AET. The estimated AET can be used for monitoring drought conditions.

[394

Standardized Vegetation Index

A standardized vegetation index (SVI) based on calculation of a z score
of NDVI distribution can also be produced for drought monitoring, as
re ported by Peters et al. (2002):
z ijk = NDVI ijk −

Line
——
12.
——
Norm
PgEn

w NDVI ij /σ ij

[30.2]

w here z ijk is z value for pixel i during week j for year k , z ijk ∼ N (0,1),

N DVI ijk is weekly NDVI value for pixel i during week j for year k , w NDVI ij
is the mean NDVI for pixel i during week j over n years, and σ ij is the stan-
da rd deviation of pixel i during week j over n years. This per-pixel proba-

[394

Figure 30.6 Monthly maximum-value-composite normalized difference vegetation indices
during 1999 for Korea. The values in the legend show NDVI values. The higher the NDVI
va lues, the better the vegetation vigor.