Pendekatan Spektral Dan Pendekatan Spasial Ekologis Untuk Estimasi Produksi Padi Menggunakan Citra Planetscope Di Kecamatan Sekampung Lampung Timur

Isnaini Dairina, Dr. Sigit Heru Murti, S.Si., M.Si

2024 | Tesis | S2 Penginderaan Jauh

Penyediaan tempat tinggal merupakan bagian Sustainable Development Goals (SDGs) nomor 1 dan 11. Namun, pengukuran kebutuhan dasar berupa tempat tinggal seringkali diabaikan atau tidak dimasukkan. Padahal permasalahan yang disebabkan oleh permukiman kota besar dan kota kecil akan memiliki dampak terhadap penerapan dan pencapaian SDGs di Indonesia Situasi ini menyebabkan diperlukan analisis hubungan backlog permukiman dengan SDGs di kota besar dan kota kecil.

Penelitian ini menggunakan Mixed Method Sequential Quan-Qual yang menggunakan metode kuantitatif seperti data sekunder indikator pencapaian SDGs dan data backlog permukiman kemudian dikombinasikan dengan data kualitatif melalui Indepth interview warga. Kemudian hubungan tersebut di analisis menggunakan SWOT untuk mengidentifikasi potensi serta tantangan dalam pembangunan kota besar Tangerang dan kota kecil Maja.

Hasil analisis menunjukkan hubungan antara pencapaian SDGs dengan backlog di kota besar Tangerang dan kota kecil Maja cenderung positif, namun tidak sepenuhnya merata. Pembangunan perumahan untuk mengurangi backlog terbukti dapat meningkatkan beberapa aspek pelaksanaan SDGs di kedua wilayah, seperti akses terhadap air bersih dan peningkatan fasilitas umum. Sebaliknya, ketika backlog terus meningkat, pelaksanaan SDGs menjadi tidak efektif, sehingga masalah perumahan seperti rumah yang tidak layak huni dapat meningkat setiap tahun. Kemudian Analisis SWOT kedua wilayah menunjukkan kota besar Tangerang memiliki kekuatan sebagai pusat ekonomi di provinsi Banten sementara kota kecil Maja dapat menjadi pusat ekonomi baru di Kabupaten Lebak. Namun kedua wilayah masih memiliki tantangan yang harus dihadapi untuk dapat memenuhi potensial tersebut.

The demand for rice is increasing as the population in Indonesia grows. This causes the need for rice production estimation data that is obtained quickly and has a high level of accuracy to provide benefits such as controlling the price of rice in the market, influencing policy making in the realm of government, and the welfare of farmers and the people. However, since 1997-2018, rice production data obtained from conventional methods have been invalid, causing a prolonged domino effect. Therefore, the application of remote sensing is needed, which has several advantages such as providing up-to-date data and covering large areas. Planetscope imagery, which has a spatial resolution of 3x3 m and a temporal resolution of 1 day, is interesting to explore considering that production estimation studies require multitemporal imagery and challenges in the form of areas with a lot of cloud cover. This study aims to (1) identify the ability of PlanetScope imagery to extract information on the use of paddy fields in the study area, (2) calculate the accuracy level of rice production estimation using the vegetation index transformation approach using PlanetScope imagery, and (3) the accuracy level of rice production estimation using the ecological spatial approach using PlanetScope imagery. The study area covers the entire Sekampung sub-district of East Lampung district, covering an area of 9,606 ha.The extraction of land use information was conducted in two experiments, namely the maximum likelihood pixel-based classification method which resulted in an overall accuracy (OA) of 74.48% (moderate agreement) and the SNIC-segmented random forest object-based classification method which resulted in an OA of 92.85% (almost perfect).The land use map generated from the object-based classification was used for further analysis.Based on the classification, growing season 1 paddy field covers 3,798 ha and growing season 2 covers 3,378 ha, which are divided into technical irrigated paddy field, double cropping rainfed paddy field, and lebak swamp paddy field.The spectral approach resulted in an estimate of 42,545.62 tonnes of harvested dry grain (GKP) or an overestimate with an accuracy rate of 89.1%.Meanwhile, the ecological approach resulted in an estimate of 37,970.71 ha tonnes of GKP or underestimate with 97.43?curacy. Both estimation results contradict the general findings because (1) the spectral approach only uses images of growing season 1 and (2) the ecological spatial approach does not have a representative sample distribution.  The concept of spectral approach production estimation that calculates productivity based on the actual condition of leaf density, rather than relying on the area of paddy fields, causes this method to be more recommended with a record of using 2x harvest season images with the consequence of producing underestimate results due to the delayed recording date of the images used due to large cloud cover in the research area. 

Kata Kunci : estimasi produksi, padi, spektral, spasial ekologis, planetscope