Conservation analysis of dipterocarps in Sarawak using geographical information systems and remote sensing
This study attempts to evaluate the status and extent of dipterocarp conservation in Sarawak as well as provides an updated quantitative information for conservation planning and identification of dipterocarp conservation hotspots. Multitemporal Landsat data of early 1990's and late 2010's...
Saved in:
Main Author: | |
---|---|
Format: | Thesis |
Language: | English English |
Published: |
2013
|
Subjects: | |
Online Access: | https://eprints.ums.edu.my/id/eprint/37663/1/24%20PAGES.pdf https://eprints.ums.edu.my/id/eprint/37663/2/FULLTEXT.pdf |
Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
id |
my-ums-ep.37663 |
---|---|
record_format |
uketd_dc |
institution |
Universiti Malaysia Sabah |
collection |
UMS Institutional Repository |
language |
English English |
topic |
SD391-410.9 Sylviculture |
spellingShingle |
SD391-410.9 Sylviculture Teo, Stephen Ping Conservation analysis of dipterocarps in Sarawak using geographical information systems and remote sensing |
description |
This study attempts to evaluate the status and extent of dipterocarp conservation in Sarawak as well as provides an updated quantitative information for conservation planning and identification of dipterocarp conservation hotspots. Multitemporal Landsat data of early 1990's and late 2010's were used for assessing spatial and temporal patterns of forest cover change in Sarawak. A total of 9 Landsat scenes covers the whole of Sarawak. As a consequence of defective scan line corrector on-board the Landsat satellite the data acquires from Landsat 7 ETM+ is suffering from missing line error since May of 2003. To overcome this problem, a series of same scene with nearest time and minimal cloud cover is chosen. The missing line error is corrected by overlaying 2 to 3 Landsat images in order to fill in the gap between the best images. Moreover, Image pre-processing is applied before overlaying. For the classification process, false colour composite of Landsat images band 5, 4, 3 to RGB is use enhance the visual interpretation where the vegetation appear as green and soil appear as red A supervised classification approach to classify the Landsat data to obtain the land cover class according to its difference in spectral signature. A post classification filtering using minority 3x3 followed by median 5x5 and were used to remove the "salt and pepper" effect. As for the land cover map on 1990, the accuracy assessment was carried out with aids of land use map which was acquired around 1990 with 1:50,000 scale which are detail enough to provide clear division of land cover classes. The accuracy of land cover map in 2009 was validated by using ground truth data and the recent hiresolution satellite preview image in from Google Earth. Ground truth verification was done along the pan-Sarawak highway and the result shows an accuracy of over 80%. It shows that forests at coastal plains were rapidly cleared for the establishment of oil palm plantations. Remaining areas of natural forest are mostly confined to protected areas and mountainous region along the Sarawak-Kalimantan border. Next, modelling the natural occurence of dipterocarp species in Sarawak is equally important for dipterocarp biodiversity conservation in order to know where they are in the landscape. Two statistical methods (Binary Logistic Regression and Multivariate Adaptive Regression Spline) and two geostatistial methods (Universal Kriging and Inverse Distance Weighting) were explored to build the best models for three selected genera (twelve species) of dipterocarps. The three genera were selected based on the whole ecological range the species in the genera cover and also include endemic and non endemic species. Herbarium distribution data were databased, georeferenced, digitized and divided into two halves for model building and validation. For statistical models, climatic, topographical and edaphic parameters were used. Proxy were used to represent the parameters which were highly correlated(p>0. 75) to avoid overfitting. Inverse Distance Weighting gives the best prediction with an average accuracy at over 80 %. Species Occurrence Models (SOMs) were then generated for all species of dipterocarps recorded in Sarawak. The species occurrence density map for each genus and category ( endemic and non endemic) was generated by overlaying the SOMs of all species in each genus or category. The species occurrence density maps were analyzed with land cover map from Landsat 7-EMT + images and protected forest areas for identifying hotspots for conservation in Sarawak. It reveals that areas in Central Sarawak and the southwest region (northwest Borneo around Kuching) are the main hotspots of dipterocarp diversity in Sarawak while the coastal lowland areas in Lower Rejang and Baram River which.are mainly peat swamp forest are poorer in species occurrence density. The difference in the hotspots is only the extent in size. In terms of endemism, as with dipterocarp diversity, the mixed diptercarp forest of central Sarawak is also the most important hotspot. Similarly the different genera also share more or less the same hotspots in the southwest region and central Sarawak. The hotspot in central Sarawak could be attributed to the overlapping of three biogeographical region-Riau pocket, northeast Borneo hotspot and biogeographical region of central Sarawak flora. The northwest Borneo around Kuching can be attributed to the heterogeneous geological makeup and older geological history. Gap analysis revealed that most protected forest areas are concentrated in southwest Sarawak ( e.g. Baka, Ku bah, Tanjung Datu and Gunung Gading National Parks) and in the northern part of Sarawak ( e.g. Niah, Lambir Hills and Mt Mulu National Parks). This leaves the hotspots in the central part of Sarawak least protected. Existing Protected areas only covers between 1.5% to 4% of the total areas for the different hotspots for all the different genera, endemic species and non endemic species. while majority of the hotspots that are still forested are outside the protected areas. |
format |
Thesis |
qualification_name |
Doctor of Philosophy (PhD.) |
qualification_level |
Doctorate |
author |
Teo, Stephen Ping |
author_facet |
Teo, Stephen Ping |
author_sort |
Teo, Stephen Ping |
title |
Conservation analysis of dipterocarps in Sarawak using geographical information systems and remote sensing |
title_short |
Conservation analysis of dipterocarps in Sarawak using geographical information systems and remote sensing |
title_full |
Conservation analysis of dipterocarps in Sarawak using geographical information systems and remote sensing |
title_fullStr |
Conservation analysis of dipterocarps in Sarawak using geographical information systems and remote sensing |
title_full_unstemmed |
Conservation analysis of dipterocarps in Sarawak using geographical information systems and remote sensing |
title_sort |
conservation analysis of dipterocarps in sarawak using geographical information systems and remote sensing |
granting_institution |
Universiti Malaysia Sabah |
granting_department |
Sekolah Perhutanan Tropika Antarabangsa |
publishDate |
2013 |
url |
https://eprints.ums.edu.my/id/eprint/37663/1/24%20PAGES.pdf https://eprints.ums.edu.my/id/eprint/37663/2/FULLTEXT.pdf |
_version_ |
1794022677155938304 |
spelling |
my-ums-ep.376632024-01-25T08:27:38Z Conservation analysis of dipterocarps in Sarawak using geographical information systems and remote sensing 2013 Teo, Stephen Ping SD391-410.9 Sylviculture This study attempts to evaluate the status and extent of dipterocarp conservation in Sarawak as well as provides an updated quantitative information for conservation planning and identification of dipterocarp conservation hotspots. Multitemporal Landsat data of early 1990's and late 2010's were used for assessing spatial and temporal patterns of forest cover change in Sarawak. A total of 9 Landsat scenes covers the whole of Sarawak. As a consequence of defective scan line corrector on-board the Landsat satellite the data acquires from Landsat 7 ETM+ is suffering from missing line error since May of 2003. To overcome this problem, a series of same scene with nearest time and minimal cloud cover is chosen. The missing line error is corrected by overlaying 2 to 3 Landsat images in order to fill in the gap between the best images. Moreover, Image pre-processing is applied before overlaying. For the classification process, false colour composite of Landsat images band 5, 4, 3 to RGB is use enhance the visual interpretation where the vegetation appear as green and soil appear as red A supervised classification approach to classify the Landsat data to obtain the land cover class according to its difference in spectral signature. A post classification filtering using minority 3x3 followed by median 5x5 and were used to remove the "salt and pepper" effect. As for the land cover map on 1990, the accuracy assessment was carried out with aids of land use map which was acquired around 1990 with 1:50,000 scale which are detail enough to provide clear division of land cover classes. The accuracy of land cover map in 2009 was validated by using ground truth data and the recent hiresolution satellite preview image in from Google Earth. Ground truth verification was done along the pan-Sarawak highway and the result shows an accuracy of over 80%. It shows that forests at coastal plains were rapidly cleared for the establishment of oil palm plantations. Remaining areas of natural forest are mostly confined to protected areas and mountainous region along the Sarawak-Kalimantan border. Next, modelling the natural occurence of dipterocarp species in Sarawak is equally important for dipterocarp biodiversity conservation in order to know where they are in the landscape. Two statistical methods (Binary Logistic Regression and Multivariate Adaptive Regression Spline) and two geostatistial methods (Universal Kriging and Inverse Distance Weighting) were explored to build the best models for three selected genera (twelve species) of dipterocarps. The three genera were selected based on the whole ecological range the species in the genera cover and also include endemic and non endemic species. Herbarium distribution data were databased, georeferenced, digitized and divided into two halves for model building and validation. For statistical models, climatic, topographical and edaphic parameters were used. Proxy were used to represent the parameters which were highly correlated(p>0. 75) to avoid overfitting. Inverse Distance Weighting gives the best prediction with an average accuracy at over 80 %. Species Occurrence Models (SOMs) were then generated for all species of dipterocarps recorded in Sarawak. The species occurrence density map for each genus and category ( endemic and non endemic) was generated by overlaying the SOMs of all species in each genus or category. The species occurrence density maps were analyzed with land cover map from Landsat 7-EMT + images and protected forest areas for identifying hotspots for conservation in Sarawak. It reveals that areas in Central Sarawak and the southwest region (northwest Borneo around Kuching) are the main hotspots of dipterocarp diversity in Sarawak while the coastal lowland areas in Lower Rejang and Baram River which.are mainly peat swamp forest are poorer in species occurrence density. The difference in the hotspots is only the extent in size. In terms of endemism, as with dipterocarp diversity, the mixed diptercarp forest of central Sarawak is also the most important hotspot. Similarly the different genera also share more or less the same hotspots in the southwest region and central Sarawak. The hotspot in central Sarawak could be attributed to the overlapping of three biogeographical region-Riau pocket, northeast Borneo hotspot and biogeographical region of central Sarawak flora. The northwest Borneo around Kuching can be attributed to the heterogeneous geological makeup and older geological history. Gap analysis revealed that most protected forest areas are concentrated in southwest Sarawak ( e.g. Baka, Ku bah, Tanjung Datu and Gunung Gading National Parks) and in the northern part of Sarawak ( e.g. Niah, Lambir Hills and Mt Mulu National Parks). This leaves the hotspots in the central part of Sarawak least protected. Existing Protected areas only covers between 1.5% to 4% of the total areas for the different hotspots for all the different genera, endemic species and non endemic species. while majority of the hotspots that are still forested are outside the protected areas. 2013 Thesis https://eprints.ums.edu.my/id/eprint/37663/ https://eprints.ums.edu.my/id/eprint/37663/1/24%20PAGES.pdf text en public https://eprints.ums.edu.my/id/eprint/37663/2/FULLTEXT.pdf text en validuser dphil doctoral Universiti Malaysia Sabah Sekolah Perhutanan Tropika Antarabangsa |