Scholarlink

<p>Daily rainfall impacts depend on sequencing and persistence as well as seasonal totals, particularly in monsoon regions where wet spells and dry breaks govern antecedent wetness and water stress. We characterize daily rainfall organization over Odisha using a multi-state hurdle-Gamma hidden Markov model (HMM) applied to India Meteorological Department (IMD) 0.25° gridded rainfall (1901–2024) across ten agroclimatic zones, analyzed separately for the monsoon (JJAS) and post-monsoon (OND) seasons. Bayesian Information Criterion (BIC) selection consistently supports a four-regime structure (K = 4) in all zones and both seasons, robust to threshold and period sensitivity checks. The inferred regimes form an intensity-ordered system (Dry, Light Wet, Wet, Very Wet) with strong seasonal contrasts in persistence. In JJAS, all regimes persist on multi-day timescales (median expected durations ~ 2.9–3.6 days), whereas OND is dominated by a highly persistent Dry background (median expected duration 18.65 days) punctuated by short wet-type episodes (~ 2.5–2.9 days). Mean regime occupancy shows coherent zone-scale organization, largely expressed as redistribution among wet-intensity subclasses rather than a simple dry–wet shift. Climate associations are likewise expressed as regime reallocation: in JJAS, the Niño3.4 index is linked to increased Dry occupancy (+ 0.037 per + 1 SD) and reduced Wet occupancy (− 0.035 per + 1 SD), while Bay of Bengal SST anomalies primarily redistribute time within wet regimes. In OND, Bay of Bengal SST anomalies reduce Dry occupancy (− 0.021 per + 1 SD) and increase Light Wet and Wet occupancy (+ 0.018 and + 0.013 per + 1 SD), alongside decreased Very Wet (− 0.010 per + 1 SD). Century-scale trends are modest and concentrated in Light Wet, increasing at + 0.00275 per decade (JJAS) and + 0.00237 per decade (OND). Overall, the regime framework provides a compact, sequence-aware description of Odisha rainfall that quantifies persistence, resolves spatial organization, and links Indo-Pacific variability to changes in rainfall-regime composition.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Agroclimatic contrasts in Odisha rainfall regimes across monsoon and post-monsoon seasons using a hurdle-Gamma hidden Markov model

  • 期刊论文

  • 发表日期：

    2026年4月10日

查看全文
  • Tapas Ranjan Patra

摘要

Daily rainfall impacts depend on sequencing and persistence as well as seasonal totals, particularly in monsoon regions where wet spells and dry breaks govern antecedent wetness and water stress. We characterize daily rainfall organization over Odisha using a multi-state hurdle-Gamma hidden Markov model (HMM) applied to India Meteorological Department (IMD) 0.25° gridded rainfall (1901–2024) across ten agroclimatic zones, analyzed separately for the monsoon (JJAS) and post-monsoon (OND) seasons. Bayesian Information Criterion (BIC) selection consistently supports a four-regime structure (K = 4) in all zones and both seasons, robust to threshold and period sensitivity checks. The inferred regimes form an intensity-ordered system (Dry, Light Wet, Wet, Very Wet) with strong seasonal contrasts in persistence. In JJAS, all regimes persist on multi-day timescales (median expected durations ~ 2.9–3.6 days), whereas OND is dominated by a highly persistent Dry background (median expected duration 18.65 days) punctuated by short wet-type episodes (~ 2.5–2.9 days). Mean regime occupancy shows coherent zone-scale organization, largely expressed as redistribution among wet-intensity subclasses rather than a simple dry–wet shift. Climate associations are likewise expressed as regime reallocation: in JJAS, the Niño3.4 index is linked to increased Dry occupancy (+ 0.037 per + 1 SD) and reduced Wet occupancy (− 0.035 per + 1 SD), while Bay of Bengal SST anomalies primarily redistribute time within wet regimes. In OND, Bay of Bengal SST anomalies reduce Dry occupancy (− 0.021 per + 1 SD) and increase Light Wet and Wet occupancy (+ 0.018 and + 0.013 per + 1 SD), alongside decreased Very Wet (− 0.010 per + 1 SD). Century-scale trends are modest and concentrated in Light Wet, increasing at + 0.00275 per decade (JJAS) and + 0.00237 per decade (OND). Overall, the regime framework provides a compact, sequence-aware description of Odisha rainfall that quantifies persistence, resolves spatial organization, and links Indo-Pacific variability to changes in rainfall-regime composition.