Bogor – Bogor University of Agriculture (IPB) Professor Imas Sukaesih Sitanggang has developed an information system on patrolling to prevent the forest and land fires (Karhutla), which is called SIPPK, for Sumatra region.
The information system is developed under a partnership with the climate change and forest and land fires control agency’s (PPIKHL) Sumatra regional office and the ministry of environment and forestry (KLHK).
Imas said that the area engulfed by the forest and land fires (Karhutla) in Indonesia in 2015 reached more than 2.6 million hectares. But in 2019 such areas only amounted to around 1.6 million hectares. To reduce the negative impacts of Karhutla, the government has prioritized efforts of prevention through a system of early detection of the Karhutla.
“One indicator of Karhutla is the hotspots recorded by remote sensing satellite. The KLHK has set the procedures to check the hotspots on the ground as stipulated in the environment and forestry ministry regulation (Permen LHK) No.8/2018,” Imas said in a press statement last week.
According to her, the regulation stipulates three priorities to check the hotspots of Karhutla on the field. The three priorities are the hotspots that are clustered, the hotspots that are delineated with satellite image indicate smokes, and the hotspots that occurred for three straight days at least.
He added that mining data approach has been known since 1990s. The mining data will provide necessary information needed for the process of taking decision.
“The data on the hotspots that is registered every day will produce a large amount of information which is intriguing to analyze. Based on the Permen, one of the intriguing patterns that can be studied from the hotspots is the clustered patterns,” she said.
Imas applies the technique of spatiotemporal clustering in using the mining data on the hotspots which have spatial, non-spatial and temporal attributes to get the clusters of the hotspots. The algorithm that is used to form the hotspot clusters is the density based spatial clustering algorithm with Noise (DBSCAN) and Spatio-Temporal DBSCAN (ST-DBSCAN).
The DBSCAN algorithm works well on the spatiality to classify objects based on density. The implementation of DBSCAN algorithm on the 2013 data of hotspots produces clusters of hotspots with highest density at 0.5 per square kilometer. The clustered hotspots generally occur on peatlands with the type of ‘Hemists/Saprists 2 (60/40), which has the depth of more than 400 cm and mostly scattered across the regencies of Rokan Hilir, Bengkalis, Dumai and Siak in Riau province.
The next temporal aspect was then involved in the process of clustering the 2015 and 2019 hotspots by using ST-DBSCAN algorithm. The result shows that around 50.8 percent and 50.3 percent of hotspots formed clusters in Sumatra, consecutively in 2015 and 2019. Based on the results, the teams of Karhutla prevention patrol are suggested to give priority to the hotspots for field verification.
“The implementation of taking the clustering patterns from the dataset of the hotspots will be proposed to be integrated with the SIPPK in the future. It is aimed to allow the Karhutla prevention patrol team to get the information on the hotspots that should be prioritized for field checking, so that efforts of preventing Karhutla can be implemented more efficiently,” she said. (*)
Source: medcom.id | Picture via mainmain.com
