Scholarlink

<p>The effectiveness of a triple collocation based non-stationary data merging scheme for estimating low uncertainty Soil Moisture (SM) over India is examined here. The SM products are combined based on their uncertainty estimates, that vary both in space as well as in time. The scheme is tested for four contemporary Land Surface Models (LSMs): NOAH, NOAH-MP, MOSAIC, and CLSM. The merging is achieved by integrating the spatiotemporal uncertainties from the LSMs with multi-satellite merged active (CCI-A) and passive (CCI-P) SM products. The evaluation of merging efficacy is achieved using the cross-correlation estimates (correlation between estimated SM and unknown true SM) from extended triple collocation (ETC) approach. The study highlighted that NOAH and NOAH-MP had relatively higher initial correlation compared to MOSAIC and CLSM. The merged SM products showed significant improvement in cross-correlation, relative to their parent LSMs, with a median improvement of 12.1% in NOAH, 16.1% in NOAH-MP, 104.4% in MOSAIC, and 39.9% in CLSM. The improvement is especially evident in the central and western India, regions with higher spatiotemporal variability of SM. The merged SM also showed higher cross-correlation relative to a contemporary stationary multi-satellite merged SM: all-active + all-passive, merged SM (CCI-C). A comparison between the non-stationary and the traditional stationary merged SM revealed greater improvement by the non-stationary scheme relative to the stationary one. This revealed the efficacy of the non-stationary merging scheme, particularly in regions with higher spatiotemporal variability, compared to stationary merging schemes.</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

Evaluating a non-stationary merging framework for improved multi-source soil moisture over India

  • 期刊论文

  • 发表日期：

    2026年7月11日

查看全文
  • Aniket Chakravorty,
  • Gayathridevi Salila,
  • Rekha Bharali Gogoi,
  • Shayam Sundar Kundu,
  • Shiv Prasad Aggarwal

摘要

The effectiveness of a triple collocation based non-stationary data merging scheme for estimating low uncertainty Soil Moisture (SM) over India is examined here. The SM products are combined based on their uncertainty estimates, that vary both in space as well as in time. The scheme is tested for four contemporary Land Surface Models (LSMs): NOAH, NOAH-MP, MOSAIC, and CLSM. The merging is achieved by integrating the spatiotemporal uncertainties from the LSMs with multi-satellite merged active (CCI-A) and passive (CCI-P) SM products. The evaluation of merging efficacy is achieved using the cross-correlation estimates (correlation between estimated SM and unknown true SM) from extended triple collocation (ETC) approach. The study highlighted that NOAH and NOAH-MP had relatively higher initial correlation compared to MOSAIC and CLSM. The merged SM products showed significant improvement in cross-correlation, relative to their parent LSMs, with a median improvement of 12.1% in NOAH, 16.1% in NOAH-MP, 104.4% in MOSAIC, and 39.9% in CLSM. The improvement is especially evident in the central and western India, regions with higher spatiotemporal variability of SM. The merged SM also showed higher cross-correlation relative to a contemporary stationary multi-satellite merged SM: all-active + all-passive, merged SM (CCI-C). A comparison between the non-stationary and the traditional stationary merged SM revealed greater improvement by the non-stationary scheme relative to the stationary one. This revealed the efficacy of the non-stationary merging scheme, particularly in regions with higher spatiotemporal variability, compared to stationary merging schemes.