NDVI - Rainy Season NDVI - Dry Season NDVI 
Regions ; 
Mean Min. Max. Mean Min. Max. Ye 
0 
Miranda 077. 020 090 06) . 016 0.89 10.87 
di 0.66  -015 088 050. —035. 033 11.44 
Negro 
Baiso 0.71 0.38 0.89 0.62 0.28 0.83 12.16 
Negro 
Planicie 0.73 0.11 0.89 063 -025 085 13.23 
Negro 
At 067 01) — 088 057. 0.04 A 0.36 13.99 
Aquid.. 
Taquari 0.67  -043 089 058. -048 . 086 14.05 
Baio 067 047" 038 057 -040 036 14.07 
Aquid.. 
Taboco 0.72 0.32 088 0.61 0.23 0.83 15.05 
Sao 0.76 0.44 087 064 031 0.85 15.64 
Lourenço 
Poconé 0.79 059 094 067 -060 088 15.79 
Entorno 066 5.052 089 054 .037 087 17.84 
Paraguai 067. 032 0% - 034 -07% 086 19.24 
Bun 066 -0.18 . 036 053 010 081 20.07 
Nheco. 
Alta 062 -015 033 048 -047 086 21.55 
Nheco. 
Câceres 0.65  -051 089 . 050  -060 082 22.44 
Paiaguäs 0.64 | -020 090 049  .040 06 23.61 
Cuiabá 060 4070 090 7052 0:7... 086 2427 
Apa 068 018 087 030 048 03 26.07 
Bardo ds 067. 029 — 051 049 019 087 26.54 
Melgaco 
Nabileque 062 -034 089 042  -049 081 32.59 
  
  
  
Table 1. NDVI mean values for each Pantanal region in rainy 
and dry seasons and the NDVI reduction between the seasons 
from the pair of MOD09GQ images 
From the 161 samples, only 63 generated good phenological 
profiles, because most of it did not have enough data to generate 
profiles in TIMESAT and some of the samples were located in 
the same pixel, so only few data could be used. Most of the 
samples were classified as group 4 - sandy samples (N=39), 
what was expected since the Pantanal sediments are mostly 
sandy. Only nine samples were classified as group 3, seven as 
group 2 and eight as group 1. 
For each group with different textural compositions, based on 
the amount of fine-grained sediments we generated a time-series 
of the mean NDVI values (Figure 3). In areas characterized 
mostly by sandy sediments (groups 3 and 4) the NDVI 
(chlorophyll) decreased quicker after the rainy season and 
varied more along the seasons, while in areas with some siltic- 
argillaceous sediments (groups 1 and 2) the NDVI decreased 
slower and had less variation. Group 1 presented a slower 
increase and decrease in chlorophyll, while the other groups 
decreased more rapidly after the end of the rainy season. The 
response of vegetation to drought effects in soils with more than 
50% of sand were very similar (groups 3 and 4), showing a 
rapid and intense decrease in chlorophyll (NDVI). The standard 
deviation (SD) is lower in group 1, but did not follow any 
patterns in the other groups. 
International Archives of the Photogrammetry, Remote Sensing and Spatial Information Sciences, Volume XXXIX-B8, 2012 
XXII ISPRS Congress, 25 August — 01 September 2012, Melbourne, Australia 
  
3600 i 
7000 i 
NDVI 
6000 
  
  
5000 
  
  
  
  
  
4000. 
  
Figure 3. NDVI time series representing the mean values for 
each group with different textural compositions 
The seasonality parameters for each group can be seen in Table 
2 and Figure 4. It can be easily verified that these data confirm 
the hypothesis. The range of variation is perfectly correlated 
with granulometric classes: the finer the sediment, the lower the 
seasonal variability. This statement, which can now be done, 
was the basis of the work hypothesis. And really, during the 
droughts the vegetation response comes to the lowest degree in 
the proposed classes. At the rainy season, with water 
disponibility, high temperatures and the graminea presence, the 
Group 4 presents a maximum NDVI response discretely higher 
than the other classes. On the other side, the minimum is hardly 
affected by the drought and the NDVI response is near perfectly 
correlated with the pelites presence. The mean NDVI response 
is, naturally, highly correlated with the pelites. The fact that 
Group 4, the more sandy, does not have a smaller response than 
Group 3 may be explained by the presence of graminea, often 
presenting planophyll leaves which are highly reflective in the 
NIR band. Finally, the standard deviation (SD) is not very 
important and is more discrete for the pelitic class, as expected. 
The reading of all the dataset used here shows a high percentage 
of agreement between the behavior of the temporal series and 
the punctual behavior of vegetation (NDVI), what can also be 
seen in Figure 2. This allied to fieldwork allows to evaluate the 
dominant particles size in areas of difficult access and well 
defined seasons. These results makes possible to proceed at the 
core of this research project: propose a subdivision of the 
Pantanal in geological and environmental homologous areas. 
  
  
  
  
Seasonality 
Parameters Groupl  Group2 Group 3 Group 4 
Mean 
Lin 17.69 18.74 18.58 18.09 
MeanBase 7010,33. 5974.79... 5022-08 511040 
value 
Mean Poa 60.21 60.46 60.76 60.28 
time 
Mean 
13.29 
Tm 1604.57 2298.43 261023 26 
Maximum 9018.30 | 8991.63 9122.44 9323.46 
Minimum 5250.92 4228.76 3542.16 2894.51 
Mean 7862.23 7138.43 6408.97 6453.02 
SD 672.46 1142.51 1067.19 1120.87 
  
Range 3767.38 4762.87 5580.28 6428.95 
  
Table 2. Seasonality parameters for each group with different 
textural compositions 
  
  
  
   
  
   
      
      
  
  
  
  
      
    
   
    
     
      
    
   
       
       
     
   
      
    
   
    
    
   
    
   
   
   
   
   
   
   
   
   
   
    
    
  
    
  
     
      
     
    
    
      
  
   
Mit 
co 
ND 
clea 
rem 
and 
sam 
anni 
the 
area 
of N 
ton 
This 
pro 
grat 
drai 
sedi 
whe 
san 
pre: 
low 
resc 
esp: 
rain 
Hig 
pho 
whe 
refl 
and 
et: 
hig 
and 
het 
text 
atte 
Pro 
inte 
the: 
Wa 
cor 
thai