農学における情報利用研究フォーラムグループ > 農学における情報利用ゼミナール

National Forest Monitoring System of Indonesia

poster(PDF)

日時：	2018年4月24日（火）16:00〜
教室：	農学部7号館A棟7階会議室
協賛:	日本農業気象学会 リモートセンシング・GIS研究部会 農業情報学会 リモートセンシング部会
参加申込・問合先：	細井文樹（生物環境情報工学） mail: ahosoi[at]mail.ecc.u-tokyo.ac.jp

abstract:
To monitor the condition of forests, mainly deforestation, since 1990 the government of Indonesia has conducted a visual classification of land cover based on Landsat satellite imagery at a scale of 1 : 250 000. This classification resulted in 23 classes of land cover. At the beginning, data throughout Indonesia were published every 3 years, then in 2010 published every two years and since 2012 were published every year. The deforestation that occurred in the period 2013-2014 amounted to 0.4 million ha/year in both forest and non-forest areas (Directorate of Inventory and Utilization of Forest Resources, 2015).
At the global level have also been carried out forest mapping using digital classification using Landsat satellite imagery (Hansen et al., 2013). By using this data Indonesian deforestation in the period 2011 – 2012 reach 2 million ha / year. This figure is more than double compare to the figures published by the Ministry of Forestry. Huge differences in the calculation have made the accountability of national data published officially by the Indonesian government decrease. Actually, this is due to the classification methodology differences and definition of the forest. The differences lead to the confusion of the implementation of REDD and calculation of the global reduction target of greenhouse gas emissions from forestry and land use sectors.
Forest definition required by the United Nations Framework Convention on Climate Change (UNFCCC) includes three variables, namely the percentage crown (crown cover), High tree (tree height) and the minimum area (minimum area) (http: //cdm.unfccc. int / DNA / index.html). Indonesia has assigned the definition of forest are: a minimum area should be larger than 0.5 ha, with a minimum height of 5 m and a canopy cover greater than 30%. The definition of forest area by using the minimum criteria can be determined easily, but if also incorporate criteria for the classification of high and canopy forests using satellite imagery data digitally need a series of further research. We need to develope a reliable method that can provide information of forest cover and forest cover changes immediately through integration of the multi-temporal imagery and other remote sensing datasets, such as the SAR data, LIDAR, DSM/DTM and high-resolution optical imageries.
Moreover, regarding to the deforestation and forest degradation issue in Indonesia, it is an urgent need to develop near-real time forest monitoring, that can be seen by the public to make sure the transparency of forest resources management. Remote sensing technology seems to be a powerful tool for monitoring and assessing the changes in forest cover immediately. Our results indicated that the use of MODIS data as a basic information in the near-real time system offers great promise to detect the forest cover change in Indonesia’s forestland, since about 90.07% of area assigned to be a change area; had actually changed. Meanwhile, about 34.97% of the forest cover change was not assigned to be a change area by the system. This result showed the need to evaluate the threshold in detecting forest cover change. Although the threshold issue is quite problematic in forest cover change detection, the results show that the methodology proposed in this study provides sufficient and useful information in forest monitoring; this includes the location, time and trajectories of the changes.