Pulau Tanakeke merupakan salah satu pulau dengan hutan mangrove yang luas di pesisir Sulawesi Selatan. Hutan mangrove ini menjadi ekosistem penting bagi masyarakat sekitar karena nilai ekologi maupun ekonominya. Namun, dalam kurun waktu sekitar tahun 1980-2000, keberadaan mangrove tersebut terancam oleh perubahan penggunaan lahan dan juga pemanfaatan yang berlebihan. Penelitian ini bertujuan untuk menganalisis perubahan temporal luas dan tingkat kerapatan hutan mangrove di Pulau Tanakeke antara tahun 2016 dan 2019. Metode analisis perubahan luasan hutan mangrove menggunakan data citra satelit Sentinel-2 multi temporal berdasarkan hasil klasifikasi hutan mangrove dengan menggunakan random forest pada platform Google Earth Engine. Akurasi keseluruhan hasil klasifikasi hutan mangrove tahun 2016 dan 2019 sebesar 91% dan 98%. Berdasarkan hasil analisis spasial diperoleh perubahan penurunan luasan mangrove yang signifikan dari 800,21 ha menjadi 640,15 ha. Kerapatan mangrove di Pulau Tanakeke sebagian besar tergolong kategori dalam kerapatan tinggi.
The topographic effect on satellite imagery has long been acknowledged and several methods have been proposed to address it. These methods mostly employ a digital elevation model to identify topographic conditions. The availability of various digital elevation models (DEMs) with different spatial resolutions prompts a thorough investigation to select suitable data for use when correcting the topographic effect on high-resolution satellite imagery. The release of Digital Elevation Model Nasional (DEMNAS) with its 8-meter spatial resolution provides a similar spatial resolution with SPOT-6/7 multispectral data (6 meters). This study presents our results for topographic correction performed using three different DEMs on orthorectified SPOT-6/7 multispectral data. These DEMs are Shuttle Radar Topography Mission (SRTM) and ALOS World 3D 30 meters (AW3D30), as well as DEMNAS. All three DEMs were resampled to match SPOT-6/7 spatial resolution (6 meters). Atmospheric correction using the MODTRAN-4 algorithm was conducted on the SPOT-6/7 multispectral images. Our study was conducted on two test sites located in the mountainous region over South Sulawesi Province, Indonesia. The Minnaert correction was chosen as the correction algorithm with the k constant calculated for each band over forest land cover. To evaluate the performance of each DEM, visual evaluation and statistical assessment were employed. Pixel values before and after topographic correction were compared over sunlit as well as shaded forest. Coefficient of variation (CV) was used as the statistical assessment tool. Our results show that AW3D30 is able to reduce the topographic effect on SPOT-6/7 multispectral images. The correlation (r) between image surface reflectance value and local illumination were reduced from 0.78 to -0.06 for the best performer on the NIR infrared band. CV was also reduced from 24.46 to 19.02 for the same NIR band. AW3D30 performed the best without the apparent under-and over-correction produced by the two other DEMs. Tweaks and modifications are found to be necessary to resolve the under-correction encountered when using SRTM and the over-correction associated with using DEMNAS on SPOT-6/7 multispectral imagery.
Spatial distribution and concentration of Total Suspended Solid (TSS) is one of the coastal parameters which are required to be examined in order to understand the quality of the water. Rapid development of remote sensing technology has resulted in the emergence of various methods to estimate TSS concentration. SPOT-6 data has spatial, spectral, and temporal characteristics that can be used to estimate TSS concentration. The purposes of this research are (1) to determine the best method for estimating TSS concentration, (2) to map TSS distribution, and (3) to determine the correlation between TSS concentration and chlorophyll-a concentration using SPOT-6 data in Segara Anakan. The estimation of TSS concentration in this research was performed using empirical model built from SPOT-6 and TSS field data. Bands used in this research are single band data (blue, green, red, and near infrared) and transformed bands such as band ratio (12 combinations), Normalized Difference Suspended Solid Index (NDSSI), and Suspended Solid Concentration Index (SSC). The result shows that blue, green, red, and near infrared bands and SSC index significantly correlated to TSS. Afterwards, regression analysis was performed to determine the function that can be used to predict TSS concentration using SPOT-6 data. Regression function used are linear and non-linear (exponential, logarithmic, 2nd order polynomial, and power). The best model was chosen based on the accuracy assessment using Standard Error of Estimate (SE). The selected model was used to calculate total TSS concentration and was correlated with chlorophyll-a field data. The result of accuracy test shows that the model from blue band has an accuracy of 70.68 %, green band 70.68 %, red band 75.73 %, near infrared band 65.58 %, and SSC 73.67 %. The accuracy test shows that red band produced the best prediction model for mapping TSS concentration distribution. The total TSS concentration, which was calculated using red band empirical model, is estimated to be 6.13 t. According to the correlation test, TSS concentration in Segara Anakan has no significant correlation with chlorophyll-a concentration, with a coefficient correlation value of -0.265.
Machine learning (ML) algorithms are frequently used in landslide susceptibility modeling. Different data handling strategies may generate variations in landslide susceptibility modeling, even when using the same ML algorithm. This research aims to compare the combinations of inventory data handling, cross validation (CV), and hyperparameter tuning strategies to generate landslide susceptibility maps. The results are expected to provide a general strategy for landslide susceptibility modeling using ML techniques. The authors employed eight landslide inventory data handling scenarios to convert a landslide polygon into a landslide point, i.e., the landslide point is located on the toe (minimum height), on the scarp (maximum height), at the center of the landslide, randomly inside the polygon (1 point), randomly inside the polygon (3 points), randomly inside the polygon (5 points), randomly inside the polygon (10 points), and 15 m grid sampling. Random forest models using CV–nonspatial hyperparameter tuning, spatial CV–nonspatial hyperparameter tuning, and spatial CV–forward feature selection–no hyperparameter tuning were applied for each data handling strategy. The combination generated 24 random forest ML workflows, which are applied using a complete inventory of 743 landslides triggered by Tropical Cyclone Cempaka 2017 in Pacitan Regency, Indonesia, and 11 landslide controlling factors. The results show that grid sampling with spatial CV and spatial hyperparameter tuning is favorable because the strategy can minimize overfitting, generate a relatively high-performance predictive model, and reduce the appearance of susceptibility artifacts in the landslide area. Careful data inventory handling, CV, and hyperparameter tuning strategies should be considered in landslide susceptibility modeling to increase the applicability of landslide susceptibility maps in practical application.
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.
hi@scite.ai
10624 S. Eastern Ave., Ste. A-614
Henderson, NV 89052, USA
Copyright © 2024 scite LLC. All rights reserved.
Made with 💙 for researchers
Part of the Research Solutions Family.