Scholarlink

<p>As the MODIS mission nears its end, ensuring continuity between legacy burned area (BA) records and products from newer sensors is essential for climate applications. Here we present the <i>Medium Resolution Burned Area Grid product version 6.0</i> (<i>MRBA60</i>), a continuous global BA time series for 2003-present, assembled on a monthly 0.25° grid by combining two components archived in separate Centre for Environmental Data Analysis (CEDA) entries: the harmonised historical MRBA60H component for 2003–2018 and the FireCCIS311 grid product from 2019 onwards. For 2003–2018, we train biome- and season-specific Random Forest models on the overlap period (2019–2024) between FireCCI51 (based on MODIS 250 m data) and FireCCIS311 (derived from Sentinel-3 SYN 300 m data) BA using auxiliary predictors (active fires, fire radiative power, climate, and vegetation indices), followed by a quantile-mapping bias correction to fit the FireCCIS311 BA distribution better. MRBA60 increases global BA relative to FireCCI51 by ~1.1 Mkm² of additional BA per year while preserving the reported decline in BA over the past two decades. Validation against Landsat-based and regional high-resolution products generally shows improved agreement compared with FireCCI51 and NASA’s MCD64A1.</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

MRBA60: Long-term consistent global burned area dataset from Sentinel-3 and MODIS products

  • 期刊论文

  • 出版模式：

    开放获取

  • 发表日期：

    2026年7月25日

查看全文
  • Miguel Ángel Torres-Vázquez,
  • Carlota Segura-Garcia,
  • Marco Turco,
  • Amin Khairoun,
  • M. Lucrecia Pettinari,
  • Erika Solano-Romero,
  • Mariano García,
  • Patricia Oliva,
  • Matthew W. Jones,
  • Emilio Chuvieco

摘要

As the MODIS mission nears its end, ensuring continuity between legacy burned area (BA) records and products from newer sensors is essential for climate applications. Here we present the Medium Resolution Burned Area Grid product version 6.0 (MRBA60), a continuous global BA time series for 2003-present, assembled on a monthly 0.25° grid by combining two components archived in separate Centre for Environmental Data Analysis (CEDA) entries: the harmonised historical MRBA60H component for 2003–2018 and the FireCCIS311 grid product from 2019 onwards. For 2003–2018, we train biome- and season-specific Random Forest models on the overlap period (2019–2024) between FireCCI51 (based on MODIS 250 m data) and FireCCIS311 (derived from Sentinel-3 SYN 300 m data) BA using auxiliary predictors (active fires, fire radiative power, climate, and vegetation indices), followed by a quantile-mapping bias correction to fit the FireCCIS311 BA distribution better. MRBA60 increases global BA relative to FireCCI51 by ~1.1 Mkm² of additional BA per year while preserving the reported decline in BA over the past two decades. Validation against Landsat-based and regional high-resolution products generally shows improved agreement compared with FireCCI51 and NASA’s MCD64A1.