Scholarlink

<p>Accurate weather forecasting is critical for addressing the escalating impacts of climate change, particularly in vulnerable coastal regions. This study focuses on Visakhapatnam, India, which faces significant risks from intensifying heatwaves, cyclones, extreme monsoons and rising sea levels. Using the Seasonal Autoregressive Integrated Moving Average (SARIMA) model, 30&#xa0;years of weather data pertaining to 1990–2019 were analysed to forecast trends for 2019–2028, with parameters including temperature, rainfall and relative humidity. Results revealed cyclic variations across all parameters, strongly influenced by urbanization and topography. Summer forecasts indicated peak maximum temperatures of 35.6–37.0&#xa0;°C and minimum temperatures rising to 27.1–28.1&#xa0;°C by 2026, highlighting the urban heat island effect. Rainfall showed high variability with predicted maxima of 135–158&#xa0;mm in the rainy season (2027), while relative humidity was projected to stabilize between 74 and 78% by 2028. Model accuracy was rigorously validated on a hold-out test set (2015–2019) with RMSE values ranging from approximately 1.15<sup>0</sup>C (for T<sub>max</sub>) to 6.5&#xa0;mm (for Rainfall), and MAPE values between 3.5 and 7.0%. Spatial analysis using GIS highlighted hotspots in urban areas such as Gajuwaka, Airport and Steel Plant, compared to cooler rural regions like Padmanabham. These findings offer actionable insights for climate-resilient urban planning, particularly in mitigating urban heat stress and managing water resources. While limitations in parameter scope are acknowledged, this study demonstrates the transformative potential of integrating SARIMA time-series forecasting and GIS for localized seasonal weather forecasting, providing policymakers with a robust tool for resilience planning.</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

Spatio-temporal seasonal climate forecasting and vulnerability mapping in Visakhapatnam, India, using an integrated SARIMA-GIS framework

  • 期刊论文

  • 出版模式：

    开放获取

  • 发表日期：

    2025年12月29日

查看全文
  • Anju Yajjala,
  • Rajeswari Erasala,
  • Murali Krishna Gurram,
  • Siva Parvati Injam

摘要

Accurate weather forecasting is critical for addressing the escalating impacts of climate change, particularly in vulnerable coastal regions. This study focuses on Visakhapatnam, India, which faces significant risks from intensifying heatwaves, cyclones, extreme monsoons and rising sea levels. Using the Seasonal Autoregressive Integrated Moving Average (SARIMA) model, 30 years of weather data pertaining to 1990–2019 were analysed to forecast trends for 2019–2028, with parameters including temperature, rainfall and relative humidity. Results revealed cyclic variations across all parameters, strongly influenced by urbanization and topography. Summer forecasts indicated peak maximum temperatures of 35.6–37.0 °C and minimum temperatures rising to 27.1–28.1 °C by 2026, highlighting the urban heat island effect. Rainfall showed high variability with predicted maxima of 135–158 mm in the rainy season (2027), while relative humidity was projected to stabilize between 74 and 78% by 2028. Model accuracy was rigorously validated on a hold-out test set (2015–2019) with RMSE values ranging from approximately 1.150C (for Tmax) to 6.5 mm (for Rainfall), and MAPE values between 3.5 and 7.0%. Spatial analysis using GIS highlighted hotspots in urban areas such as Gajuwaka, Airport and Steel Plant, compared to cooler rural regions like Padmanabham. These findings offer actionable insights for climate-resilient urban planning, particularly in mitigating urban heat stress and managing water resources. While limitations in parameter scope are acknowledged, this study demonstrates the transformative potential of integrating SARIMA time-series forecasting and GIS for localized seasonal weather forecasting, providing policymakers with a robust tool for resilience planning.