Scholarlink

This study presents a novel methodology to forecast shoreline position at the seasonal to five-year time scale using the statistical time series models SARIMAX and SVARX in combination with satellite-derived shorelines (CoastSat). 25 years of CoastSat shorelines at Curl Curl Beach in SE Australia were used to test and train the models, with forecast inputs including offshore and nearshore significant wave height and several climate indices relevant to the Australian coastline (ENSO and IOD). Over a five-year forecast horizon, it was found that the SARMIAX and SVARX models showed higher skill than a baseline linear regression model, with RMSE of approximately 10 m across all nine beach transects. The performance of these statistical time series models indicates the potential for the methodology to be extended to large spatial scales, providing valuable long range (i.e. seasonal to multi-year) forecasts of shoreline change to assist coastal managers.

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

A Methodology for Seasonal and Long-Range Forecasting of Shoreline Erosion Hazards Across Australia

  • 会议论文

  • 出版模式：

    开放获取

  • 发表日期：

    2026年

查看全文
  • Katie Wilson,
  • Mitchell D. Harley,
  • Kristen D. Splinter

摘要

This study presents a novel methodology to forecast shoreline position at the seasonal to five-year time scale using the statistical time series models SARIMAX and SVARX in combination with satellite-derived shorelines (CoastSat). 25 years of CoastSat shorelines at Curl Curl Beach in SE Australia were used to test and train the models, with forecast inputs including offshore and nearshore significant wave height and several climate indices relevant to the Australian coastline (ENSO and IOD). Over a five-year forecast horizon, it was found that the SARMIAX and SVARX models showed higher skill than a baseline linear regression model, with RMSE of approximately 10 m across all nine beach transects. The performance of these statistical time series models indicates the potential for the methodology to be extended to large spatial scales, providing valuable long range (i.e. seasonal to multi-year) forecasts of shoreline change to assist coastal managers.