Full text: Proceedings, XXth congress (Part 2)

  
| 
I 
International Archives of the Photogrammetry, Remote Sensing and Spatial Information Sciences, Vol XXXV, Part B2. Istanbul 2004 | Interna 
  
the distinction of Pinus patula plantations and the land cover further pixels of that class are expected. A reason for the occur- covere« 
class ^Near natural old secondary forest". As Pinus patula is rence of such pixels scattered throughout the class "Near natural darv fo 
an exotic species in Kenya, apart from forest plantations no * old secondary forest" (1) could be that despite preprocessing | / shrub 
minor shading effects due to the variable terrain | Forest 
have still some influence. But also young pine ha of c 
plantation show similar spectral reflection as class | The rat 
1. Further inaccuracies are introduced in the areas 1979/8 
where masked-out clouds have been replaced by | steep s 
the image information of the second scene. Over- | change 
all, the classification results for the time step | found 2 
1994/95 is judged to be the most accurate, fol- | are not 
lowed by the time step 2001. The quality for 1989 | compar 
and 1984 is less because only one scene (dry sea- | classes. 
son) is at hand and thus monoseasonal processing 
had to be performed. The classifications of MSS- | 43 CI 
data are of poorer quality due to their coarser 
  
      
   
    
  
  
  
  
  
  
  
  
  
  
  
  
  
  
  
  
  
  
  
  
  
  
  
  
  
  
  
  
  
  
  
  
  
  
  
  
  
  
  
  
spatial resolution of 60 x 60 m^ as well as their In som 
lower number in spectral channels (e.g. mid infra- forest b 
red not covered). Within the results based on MSS- | the fore 
data the classification of 1972/73 seems to be (based 
best, followed by the reults of the time steps 1975 | digitize 
and 1979/80. In spite of drawbacks regarding ries fro 
E differentiated information for detailed studies, | buffere 
1 od secendary1orest EN MSS-data allows to get back in time as far as the | comple 
A 2 Secondary forest [77] E early 1970s. Just making use of ETM+/TM-data ensured 
E S a teten Ez would limit time series to the past 20 years. | the con 
oS hand, t 
5 S scattered trees 4.Y Areas within Official Forest Boundaries | areas w 
E 6 Grassland ; 
P Plantation forest - E g : agricult 
E] 7 — Pus patula Because the forest areas are giving priority in the | importa 
Ds Bischoffajavanica BIOT A-East project, numerical analysis is concen- project. 
E| 9. Renate trating on these area. Here, the areas within the | While 
| » Lig oe official forest boundaries as gazetted in the 1930s | bounda 
€ are assessed. For Kaimosi Forest and Maragoli | derived 
LZ. 0m Forest no administrative forest boundary is shown | for anal 
rn on the topo maps 1:50,000. Therefore they are not is asses 
je am included in this analysis. Thus. for 9 of the 11 | 
Led forest areas (see Figure 1) the areas covered by the | rE 
Figure 2: Landsat ETM+ land cover classification result of 2001 for the quitte land sn e SENSUS are derived fon a ie 
i steps. As an example, for Kakamega Forest con- | 
subset of Kakamega Forest, displayed together with the siderable:chanses in "Near matural 2-old secondare 
official forest boundary as gazetted in 1933. forest” (1) and "Secondary Forest" (2) over time | 
sa : pe j Turic > are revealed: 15,000 ha of | 
[ha] South Nandi Forest, land cover within the official boundary class 1 and 2 in 1972 (compare | 
8.000 Kokwaro, 1988: 14,300 ha in 
1972) and 12,200 ha in 2001. | 
7.000 + [| This means a 20% loss in 
= forest area over the past 30 | 
SL years. By just looking at “Near | £ 
natural + old secondary forest” = 
500) 11 L2. a similar loss is observed: d 
8,500 ha in 2001 as compared 5 
to 10.500 ha in 1972. £ 
4.000 All numbers are presented via | A 
| bar charts to visualize typical 
3.000 | trends in changes of forest | 
7 | formations over time for each | 
2.000 + ud | forest. For several of the forest 
| [| areas very pronounced op- | 
1.000 - ^ Lid | posed dynamics regarding the | 
Hd | | | land cover classes “Near natu- | 
| i : : B IER. É uli | ral + old secondary forest”, 
1972/73 1975 1979/80 — 1994/95 2001 L "Secondary. ' Forest". and | 
| [——————— LANDSAT-MS§ ——————————-]-————————————- LANDSAI-IM ———————————4 | “Bushland ; shrubs* are ob- 
| | served. For an example see the | SE 
KA remet re it BB — ere ea eec tnt 
graph for South Nandi Forest 
Figure 3: Land cover 1972 -- 2001 for South Nandi Forest. area [in ha] within official SS 
mm: Figure 3 Here. the aren 
forest boundary (for colours see Figure 2) 
468 | 
 
	        
Waiting...

Note to user

Dear user,

In response to current developments in the web technology used by the Goobi viewer, the software no longer supports your browser.

Please use one of the following browsers to display this page correctly.

Thank you.