Dinamika Dampak Spatiotemporal Perubahan Tutupan Lahan Terhadap Water Yield di Sub-DAS Brantas Hulu (2005–2025)

Hekar, Devon Leonardo (2026) Dinamika Dampak Spatiotemporal Perubahan Tutupan Lahan Terhadap Water Yield di Sub-DAS Brantas Hulu (2005–2025). Other thesis, Institut Teknologi Sepuluh Nopember.

Text 5015221078_Devon Leonardo Hekar_Final.pdf Restricted to Repository staff only Download (6MB) | Request a copy

Abstract

Alih fungsi lahan yang masif di Sub-DAS Brantas Hulu mengancam stabilitas tata air dan meningkatkan risiko bencana hidrometeorologi. Penelitian ini bertujuan untuk menganalisis dinamika spatiotemporal perubahan tutupan lahan dan dampaknya terhadap luapan air permukaan (water yield) periode 2005–2025. Pengolahan tutupan lahan memanfaatkan algoritma Multi-Kernel Convolutional Neural Network (CNN) berbasis citra Landsat, dilanjutkan estimasi hidrologi menggunakan pemodelan InVEST Annual Water Yield, serta analisis keterkaitan spasial melalui Geographically Weighted Regression (GWR) berbasis unit grid 900 m × 900 m (n = 684). Hasil klasifikasi CNN (Overall Accuracy 88,33%--93,33%) menunjukkan ekspansi kawasan terbangun paling radikal sebesar +5.692 ha (+67,38%), meningkat dari 8.448 ha (7,14%) pada tahun 2005 menjadi 14.140 ha (11,95%) pada tahun 2025. Pertumbuhan ini mengorbankan lahan pertanian yang menyusut menjadi 83.161 ha (70,27%) dan degradasi kawasan hutan hingga 14.282 ha (12,07%). Alih fungsi lahan ini memicu penurunan evapotranspirasi aktual (AET) hingga 593,08 mm/tahun dan lonjakan ratarata water yield depth dari 1.289,90 mm/tahun menjadi 1.644,87 mm/tahun. Secara akumulatif, total volume air permukaan melesat sebesar +27,57% (+182.000.000 m 3/tahun) menyentuh 842.000.000 m3 /tahun pada tahun 2025. Uji statistik spasial mengonfirmasi autokorelasi positif yang menguat (Indeks Moran's I naik dari 0,8047 menjadi 0,8303 ), membentuk klaster Hot Spot di Bumiaji dan Kota Batu. Pemodelan GWR lokal terbukti jauh lebih superior dibanding OLS global dengan nilai Adjusted R2 sebesar 0,7455 (74,55%). Analisis t-statistic ( |t| > 2,702 ) membuktikan bahwa dampak ekspansi beton terhadap lonjakan water yield di tahun 2025 telah melebur menjadi krisis hidrologis ekologis yang bersifat sistemik dan regional
==================================================================================================================================
Massive land conversion in the Upper Brantas Sub-Watershed threatens hydrological stability and escalates hydrometeorological disaster risks. This study aims to analyze the spatiotemporal dynamics of land cover change and its impact on surface water yield from 2005 to 2025. Land cover classification was processed using a Multi-Kernel Convolutional Neural Network (CNN) algorithm on Landsat imagery, followed by hydrological estimation via the InVEST Annual Water Yield model, and spatial relationship analysis using Geographically Weighted Regression (GWR) on a 900 m × 900 m grid unit (n = 684). CNN classification (Overall Accuracy 88.33%--93.33% ) revealed a radical expansion of built-up areas by +5,692 ha (+67.38%), rising from 8,448 ha (7.14%) in 2005 to 14,140 ha (11.95%) in 2025. This growth degraded agricultural land to 83,161 ha (70.27%) and forest cover to 14,282 ha ( 12.07% ). Land conversion triggered a decline in actual evapotranspiration (AET) to 593.08 mm/year and a surge in mean water yield depth from 1,289.90 mm/year to 1,644.87 mm/year. Accumulatively, total surface water yield volume spiked by +27.57% (+182,000,000 m3/year ), reaching 842,000,000 m3/year in 2025. Spatial statistics confirmed strengthening positive autocorrelation (Global Moran's I increased from 0.8047 to 0.8303), forming Hot Spot clusters in Bumiaji and Batu City. Local GWR proved far superior to global OLS, achieving an Adjusted R2 of 0.7455 (74.55%). The t-statistic analysis (|t| > 2.702) demonstrates that the impact of built-up expansion on water yield surges in 2025 has mutated into a systemic, regional-scale ecological crisis

Item Type:	Thesis (Other)
Uncontrolled Keywords:	Convolutional Neural Network, Geographically Weighted Regression, InVEST, Sub-DAS Brantas Hulu, Water Yield, Upper Brantas Sub-Watershed
Subjects:	G Geography. Anthropology. Recreation > G Geography (General) > G70.5.I4 Remote sensing
Divisions:	Faculty of Civil, Planning, and Geo Engineering (CIVPLAN) > Regional & Urban Planning > 35201-(S1) Undergraduate Thesis
Depositing User:	Devon Leonardo Hekar
Date Deposited:	05 Aug 2026 02:27
Last Modified:	05 Aug 2026 02:27
URI:	http://repository.its.ac.id/id/eprint/141391

Actions (login required)

View Item