Pengembangan Layanan Iklim Prediktif Operasional Melalui Integrasi Model Musiman ECMWF SEAS5 dan Adaptasi Model Fenologi Clisagri Pada Pertanian Padi

Abstract

Penelitian ini dilatarbelakangi oleh ancaman perubahan iklim terhadap ketahanan pangan nasional, khususnya sensitivitas padi tropis pada sawah tadah hujan yang sangat bergantung pada iklim monsun dan dipengaruhi fenomena global ENSO dan IOD. Terdapat potensi kesenjangan yang cukup dalam terkait penyediaan layanan iklim yang mudah diterjemahkan menjadi adaptasi agronomis presisi di sektor pertanian. Penelitian bertujuan mengintegrasikan model prediksi musiman ECMWF SEAS5 dengan adaptasi model fenologi Clisagri guna membangun paradigma layanan iklim prediktif operasional. Kebaruan riset terletak pada transformasi model fenologi gandum subtropis untuk diadaptasi padi tropis serta pengembangan strategi optimasi waktu tanam dan varietas berbasis risiko iklim di Indonesia. Systematic Literature Review (SLR) dengan protokol PRISMA telah dilakukan menggunakan 1.981 artikel Scopus. Hasil SLR ditemukan ketimpangan riset antara strategi adaptasi (77.4%) dan penyediaan layanan iklim (22.6%). Mayoritas riset di negara berkembang masih bersifat deskriptif retrospektif pada tahap pemahaman dasar, bukan aksi presisi. Temuan menekankan urgensi pergeseran menuju layanan iklim lanjut (advanced) yang mengintegrasikan Seasonal Climate Forecast (SCF) dengan pemodelan fenologi dinamis melalui pendekatan co-design untuk meningkatkan kapasitas operasional petani. Metode evaluasi diagnostik diterapkan pada luaran model re-forecast SEAS5 terhadap data referensi MSWEP (presipitasi) dan ERA5 (suhu) periode 1991–2020 menggunakan metode korelasi Pearson dan analisis hit/miss. Hasil menunjukkan SEAS5 sangat mumpuni mereproduksi siklus tahunan monsunal (r > 0,8), namun memiliki wet bias di wilayah pegunungan dan cold bias pada suhu maksimum di wilayah barat Indonesia. Berdasarkan seleksi bertingkat (100% Hit dan korelasi sangat kuat), ditetapkan lima wilayah yang memiliki karakteristik iklim paling dominan dapat dijelaskan SEAS5 yaitu: Rote Ndao, Kota Kupang, Kabupaten Kupang, Kepulauan Tanimbar, dan Maros. Penelitian ini melakukan transformasi struktural model Clisagri melalui re- parameterisasi kebutuhan termal Growing Degree Days (GDD) spesifik padi tropis. Menggunakan metode Algoritma Genetika (GA) dan Indeks Risiko Komposit (IRK), model mensimulasikan optimasi varietas dan pergeseran waktu tanam di Maros dan Cilacap. Di Maros, ditemukan titik ekuilibrium aman pada penanaman April (risiko 0,169) melalui strategi percepatan umur genjah saat El Niño. Di Cilacap, terdeteksi rigiditas kalender tanam akibat monsun basah persisten, sehingga solusi optimal beralih pada eksploitasi genetik menggunakan varietas berumur dalam yang mampu mereduksi risiko hingga 23%. Sintesis hasil menegaskan adanya hambatan prediksi di benua maritim Maritime Continent Prediction Barrier (MCBI) yang ditandai dengan bias sistematis model global akibat kompleksitas orografis. Kebaruan metodologis dibuktikan melalui pendekatan delineasi spasial bertingkat yang mengekstraksi wilayah dengan karakteristik iklim dominan luaran data mentah SEAS5 yang berkinerja tinggi, sehingga dapat dimanfaatkan secara operasional sebelum dilakukan koreksi bias. Integrasi ini membuat terobosan antara disiplin ilmu antara klimatologi dan agronomi sehingga menyediakan informasi terstruktur yang mengubah paradigma kesadaran iklim pasif menjadi strategi adaptasi teknis yang terukur dan antisipatif. Berdasarkan hasil yang diperoleh, dapat disimpulkan bahwa integrasi model seasonal SEAS5 dan model fenologi dinamis Clisagri termodifikasi sangat efektif dalam mengkuantifikasi risiko iklim pada fase kritis pertumbuhan padi. Rekomendasi strategis berupa adopsi konsep rekayasa algoritma model ini ke dalam sistem kalender tanam nasional. Saran penelitian lanjutan mencakup integrasi prediksi skala sub-musiman (S2S) untuk dapat mendeteksi dry spells mingguan dan penyempurnaan indikator cekaman termal berbasis suhu malam hari (Tmin) guna mengantisipasi sterilitas spikelet akibat pemanasan global.

This research is driven by the threat of climate change to national food security, particularly the sensitivity of tropical rice in rainfed fields that are highly dependent on the monsoon climate and influenced by global phenomena such as ENSO and IOD. There is a significant gap in providing climate services that can be easily translated into precise agronomical adaptations in the agricultural sector. This study aims to integrate the ECMWF SEAS5 seasonal prediction model with the adaptation of the Clisagri phenology model to establish an operational-predictive climate service paradigm. The research novelty lies in the structural transformation of the Clisagri phenology model originally designed for temperate wheat to be adapted for tropical rice, along with the development of climate risk-based planting time and variety optimization strategies in Indonesia. Utilizing a Systematic Literature Review (SLR) method with the PRISMA protocol on 1,981 Scopus-indexed articles, a sharp imbalance was identified between research on adaptation strategies (77.4%) and climate service provision (22.6%). Most research in developing countries remains trapped in descriptive- retrospective approaches at the basic understanding level rather than precision action. The findings emphasize the urgent need for a shift toward advanced climate services that integrate Seasonal Climate Forecasts (SCF) with dynamic phenology modeling through a co-design approach to enhance farmers' operational capacity. A diagnostic evaluation method was applied to the ECMWF SEAS5 re- forecast outputs against MSWEP (precipitation) and ERA5 (temperature) reference datasets for the 1991–2020 period using Pearson correlation and hit/miss analysis. The results indicate that SEAS5 is highly capable of reproducing the annual monsoonal cycle (r>0.8), although it exhibits a wet bias in mountainous regions and a cold bias in maximum temperatures across Western Indonesia. Based on a tiered selection process (100% Hit rate and very strong correlation), five regions with the most dominant climate characteristics explainable by SEAS5 were established: Rote Ndao, Kupang City, Kupang Regency, Tanimbar Islands, and Maros. The study undertook a structural transformation of the Clisagri model through the re-parameterization of thermal requirements Growing Degree Days (GDD) specific to tropical rice. Using Genetic Algorithms (GA) and the Composite Risk Index, the model simulated variety optimization and planting time shifts in Maros and Cilacap. In Maros, a safe equilibrium point was identified for April planting (risk index 0.169) through a strategy of utilizing early-maturing varieties during El Nino events. In Cilacap, planting calendar rigidity was detected due to persistent wet monsoons; thus, the optimal solution shifted toward genetic exploitation using long-duration varieties, which effectively reduced the aggregate risk by up to 23%. The synthesis of results confirms the existence of the Maritime Continent Prediction Barrier (MCPB), characterized by systematic biases in global models due to orographic complexity. Methodological novelty was demonstrated through a tiered spatial delineation approach that extracts regions where raw SEAS5 high- performance outputs can be utilized operationally before statistical bias correction. This integration represents a breakthrough bridging climatology and agronomy, providing structured information that transforms the paradigm from passive climate awareness to measurable and anticipatory technical adaptation strategies. The dissertation concludes that the integration of the SEAS5 seasonal model and the modified Clisagri dynamic phenology model is highly effective in quantifying climate risks during critical rice growth phases. Strategic recommendations include the adoption of this algorithmic engineering concept into the national Integrated Crop Calendar system. Future research suggestions include integrating sub-seasonal to seasonal (S2S) predictions to detect weekly dry spells and refining thermal stress indicators based on nighttime temperatures (Tmin) to anticipate spikelet sterility due to global warming.