Hotspot and rainfall analysis in Bombana District, Southeast Sulawesi Province

Forest and land fires are starting to occur in eastern Indonesia, such as Southeast Sulawesi, frequently. Bombana District is one of the areas in Southeast Sulawesi Province with the highest total hotspots in the last decade. This study aimed to analyze the trend of hotspot distribution and its relationship with rainfall in the 2012 – 2021 period in Bombana District, Southeast Sulawesi Province. The sources used in this study were hotspot data from MODIS Terra/Aqua satellite imagery and rainfall data from the NASA Langley Research Center (LaRC) Power, which were analyzed using descriptive and regression analysis. The results showed that the total hotspots and rainfall in Bombana District between 2012 and 2021 fluctuated. The highest hotspot peak occurred in 2015, with 343 points. The hotspots and rainfall were significantly correlated, where rainfall greatly influenced the total of hotspots by 64% in 2012–2021.


Introduction
The forest area lost due to deforestation in Indonesia reached 3,556,097.5hectares from 2013 to 2020 [1].One of the factors contributing to deforestation is forest and land fires.Forest and land fires refer to incidents of forest and land burning, both naturally and as a result of human activities, leading to environmental damage [2].In Indonesia, approximately 99% of forest and land fires are caused by human activities, while 1% is attributed to natural factors [3].Human activities that can cause these fires include slash-and-burn farming, hunting wild animals in forests, and careless recreation and camping in forests, which can result in the spread of remaining sources of fire [4].
Forest and land fires in Indonesia continue to occur every year.The largest forest and land fires occurred in Indonesia was in 1997/1998 about 11 million hectares.However, severe fires reoccurred in 2015 and the burned area was 2.6 million hectares, with the majority happening in Kalimantan and Sumatra [5].These forest and land fires cause severe impacts and high losses, necessitating efforts to control and manage them.One of the methods used for controlling forest and land fires is hotspot monitoring.Hotspot monitoring is an early detection system that utilizes remote sensing technology to rapidly identify and locate forest and land fires [6].The primary objective of hotspot monitoring is to expedite the fire extinguishing process in affected areas, enabling a timelier response to prevent the fires from spreading further.Data obtained from remote sensing has several advantages, making it highly valuable and suitable for implementation in Indonesia [7].
Forest and land fires in Indonesia are increasing in the eastern part of the country, particularly in Southeast Sulawesi Province, which has a high total fire cases and hotspots [8].Bombana District in Southeast Sulawesi has recorded the highest total hotspots.To address this issue, it is crucial to gather information on fire-prone areas based on hotspot distribution and its relation to rainfall patterns.Understanding these vulnerable areas will help authorities and stakeholders allocate resources and implement preventive measures effectively, mitigating the impact of fires on the environment, communities, and local economies.This research analyzes the distribution patterns of hotspots and rainfall and their correlation in Bombana District, Southeast Sulawesi Province, from 2012 to 2021.

Study Area
The study was conducted in Bombana District (Figure 1).Bombana Regency is the result of the expansion of Buton Regency by Law No. 29 of 2003, with a land area of 3,316.16km 2 and a water area of 11,837.32km².It is located between 4°22'59.4''-5°28'26.7''South Latitude and 121°27'46.7''-122°11'9.4''East Longitude [9].This district is divided into 22 sub-districts, with Rumbia as the district capital.Based on topographic conditions, the Bombana Regency area has wavy and hilly land in general.The climate throughout the Bombana Regency area is classified as tropical, with quite significant conditions between the dry and the rainy seasons.Rainfall in Bombana Regency is less than 2,000 mm/year, air temperature is around 21-35˚C, and air humidity is about 74 -89% [1].

Data Collection
Hotspot data were obtained from Terra and Aqua MODIS active fire products, available at all confidence levels, and downloaded from NASA's Fire Information for Resource Management System (https://firms.modaps.eosdis.nasa.gov).The administrative boundary map was obtained from the official website of Tanah Air Indonesia (https://tanahair.indonesia.go.id), while peat distribution data was acquired from the Indonesian Center for Agricultural Land Resources Research and Development (BBSDLP) (http://bbsdlp.litbang.pertanian.go.id/).Daily rainfall data from 2012 to 2021 was utilized to investigate the relationship with hotspots, obtained from the NASA Langley Research Center's (LaRC) Power Data Access Viewer (https://power.larc.nasa.gov/data-accessviewer/).

Data analysis
The data analysis used in this research consists of descriptive and regression analysis.The processed data are then subjected to descriptive analysis to provide an overview of the data or draw conclusions from the presented information.According to [9], the purpose of descriptive analysis is to transform a set of raw data into a more easily understandable form, showing clear and comprehensible information.On the other hand, regression analysis is a statistical process aimed at determining the influence of one variable on another.The present research uses a simple linear regression model, which is represented by the following equation: In a regression analysis, Y represents the dependent variable or the value to be predicted.In contrast, x represents the independent variable (time index), a is the value of Y when x=0 (intercept), and B is the average change in Y for one unit change in x (slope) [10].The decision-making criteria in hypothesis testing refer to the significance level compared to the probability value of 5% (0.05): a) If the p-value is less than 0.05 (p-value < 0.05), the independent variable influences the dependent variable.In this case, the null hypothesis (H0) is rejected, and the alternative hypothesis (H1) is accepted b) If the p-value is more significant than 0.05 (p-value > 0.05), the independent variable does not influence the dependent variable.In this case, the null hypothesis (H0) is accepted, and the alternative hypothesis (H1) is rejected.

The Hotspot Number from 2012 to 2021
A hotspot is a location on the earth's surface that has a relatively higher temperature compared to the surrounding area [7].Hotspots are believed to provide indications of forest and land fires not only because their numbers are repeated and clustered but also because the confidence level (confidence level) has a significant influence on indications of forest and land fires [11].The higher the confidence interval is, the higher the potential that the hotspot is a land or forest fire that occurred.MODIS Active Fire Product User's Guide divides three classes of trust levels, which are presented in Table 1 [12].2, 2015 was the peak with the highest hotspots, 343.These results align with the increase in forest and land fires in Indonesia.The 2015 forest and land fires were the worst during the last 10 years, where around 2.6 million hectares of the area were burned ( [10].The second-highest number of hotspots was in 2019, with 27 hotspots.Hotspots increased in both years because Indonesia experienced a strong El Nino in 2015 and a weak El Nino in 2019 [11].El Nino is a climate phenomenon characterized by increased sea surface temperatures in the Pacific Ocean, leading to warmer sea surface temperatures near the equator [12].El Nino affects the eastward movement of winds, resulting in decreased rainfall in parts of Indonesia, leading to prolonged dry seasons [13].Long drought causes organic materials such as litter to dry out, making them more susceptible to fires.2020 and 2021 had the lowest total of hotspots in the last decade.The decrease in hotspot occurrences was attributed to Southeast Sulawesi, including Bombana District, experiencing a moderate La Nina in 2020-2021 [14].La Nina is the opposite of El Nino, indicating a return to normal weather conditions after an El Nino event.

Monthly Hotspot and Rainfall Distribution at Bombana District in the Period 2012 -2021
The hotspots vary each month based on the weather and climate conditions in the area.According to Figure 4, hotspots increase in July and peak in October.The community is encouraged to increase prevention efforts and reduce activities that could trigger forest and land fires leading up to these months.Hotspots do not necessarily describe forest and land fires, but hotspots that are numerous and clustered indicate that there are forest fires in an area.However, hotspots can be used as indicators of forest fires even though they do not fully show actual fire incidents.The average monthly rainfall in Bombana Regency in 2012 -2021 began to decline in July and decreased sharply in October (Figure 5).The decrease in rainfall in Bombana Regency is in line with the hotspot increase in the same month.Rainfall is a weather element that has a high correlation with the occurrence of forest fires and is the highest factor in determining fuel accumulation [3].Low rainfall can help the fuel drying process, so the fuel becomes flammable when ignited.Rainfall in 2015 showed the lowest level during that year's period, where the annual rainfall was 1542.8 mm/year.The El Nino phenomenon 2015 caused a prolonged drought; this condition was characterized by 2015 experiencing a dry season of around 5 months.The dry season is a period where monthly rainfall is ∼125 mm [15].Rainfall in 2015 began to decline in July, namely 33.76 mm/month, and decreased sharply in August (5.23 mm/month), September (5.48 mm/month), and October (9.37 mm/month).It started to increase in November (60.47 mm/month).

Hotspots Number in Sub-District within Bombana District during 2012 -2021
Bombana Regency is divided into 22 sub-districts, with Rarowatu as the district capital.Lantari Jaya District has the highest hotspots, 465 in 2012-2021 (Figure 6).The second sub-districts sub-districts sub-districts sub-district sub-districts sub-district with the most hotspots was Mata Usu, with 218 hotspots, followed by the Poleang Barat sub-district, with 184 hotspots.The high number of hotspots in the Lantari Jaya sub-district was caused by human activity.A gold mine was discovered in Bombana Regency in 2008, which has caught much attention from many parties.Lantari Jaya District is a district that is heavily impacted by mining activities, while the majority of its people depend on the agricultural sector for livelihood.The impact of gold mining activities in Lantari Jaya District is that the land becomes drier, damage to irrigation as a source of water for farmers' fields and mounds of excavated soil.The source of irrigation for farmers' rice fields in this district is a rain tank supported by a simple irrigation channel originating from Langkowala Dam.Since the mining company has dammed the water upstream for mining activities, the Langkowala Dam no longer functions.This condition caused the community's rice production to decline so much that some people lost their livelihoods and returned to farming or partnering with mining companies [16].People who returned to farming prepared land by burning forests and land because they had no funds and no choice but to burn [17].Based on the results of this analysis, it can be said that Lantari Jaya in Bombana Regency is a subdistrict prone to fire.Thus, the Lantari Jaya sub-district must prioritize efforts to control forest and land fires by the local government; apart from that, the Mata Usu sub-district and Poleang Barat sub-district need to increase their prevention efforts due to the high number of hotspots between 2012 and 2021.

Trend of hotspots number and rainfall in Bombana District the period 2012 -2021
Figure 7 shows the distribution of hotspots and monthly rainfall in Bombana District.Pada Figure 7 shows the distribution of hotspots and monthly rainfall in Bombana District.In the 2012-2021 period, hotspots in the Bombana district tended to start increasing in July and reached their peak in October.That condition was accompanied by a decrease in rainfall in July, decreasing sharply in October.During that period, the average hotspot began to decline in November, accompanied by increased rainfall.Overall, it can be said that when rainfall decreases, the number of hotspots increases, and vice versa.The number of hotspots in 2015 began to appear in July, increased sharply in September (76 hotspots), and reached the highest in October (141 hotspots).The increase in the number of hotspots is inversely proportional to monthly rainfall, where in July, rainfall starts to decrease (33.76 mm/month), drops sharply in August (5.25 mm/month), and increases again in November.There is a strong correlation between rainfall and the number of hotspots from MODIS Terra/Aqua and NOAA satellite data [18].In 2015, Bombana Regency experienced a dry season from July to November.

The Relationship between Hotspot Number and Rainfall
Based on the results of linear regression analysis, the equation model between hotspots and annual rainfall shows significant results (P value = 0.005).The hotspot regression coefficient value is negative (-0.198), indicating that the influence of the rainfall variable on the number of hotspots is inversely proportional.This inverse relationship means that a decrease will follow an increase in rainfall in the number of hotspots and vice versa.The graphic pattern connects the number of hotspots and annual rainfall (Figure 8).The R-square value in the image above is 0.64, so it is categorized as having a moderate regression line.These results mean that rainfall between 2012 and 2021 in Bombana Regency only affected 64% of the number of hotspots found, while 36% of the number of hotspots located in the research location was influenced by other factors, one of which was human activity, such as clearing and burning.

Conclusion
The highest number of hotspots between 2012 and 2021 in Bombana Regency occurred in 2015 and 2019, with 343 and 271 hotspots, respectively.In that period, hotspots increased in July and peaked in October.This condition is inversely proportional to rainfall, which started to decline in July and decreased sharply until October.The relationship between annual rainfall and hotspots was significant (P value = 0.005) with R2 (64%).Prevention efforts must begin to be increased towards July, because in this month rainfall decreases and hotspot increase.
Based on hotspot analysis between 2012 and 2021 period, Bombana is a district that is quite prone to forest and land fires in Southeast Sulawesi Province.Thus, it needs to receive more attention to control forest and land fires.Lantari Jaya is a fire-prone sub-district in Bombana Regency, so prevention efforts need to be prioritized.

Figure 6 .
Figure 6.Total of hotspots per sub-district during the period from 2012 to 2021 in Bombana District

Figure 7 .
Figure 7.The relationship between hotspots and rainfall in Bombana District in The Period 2012-2021

Figure 8 .
Figure 8. Relationship between hotspots and rainfall in Bombana District

Table 1 .
Meaning of confidence interval