Distilling Before Refine: Spatio-Temporal Transfer Learning for Mapping Irrigated Areas Using Sentinel-1 Time Series - Archive ouverte HAL Accéder directement au contenu
Article Dans Une Revue IEEE Geoscience and Remote Sensing Letters Année : 2020

Distilling Before Refine: Spatio-Temporal Transfer Learning for Mapping Irrigated Areas Using Sentinel-1 Time Series

Résumé

In agricultural areas, the downstream flow can be highly influenced by human activities during low flow periods, especially dam releases and irrigation withdrawals. Irrigation is indeed the major use of freshwater in the world. This study aims at precisely taking these factors into account in a watershed model. The Soil and Water Assessment Tool (SWAT+) agro-hydrological model was chosen for its capacity to model crop dynamics and management. Two different crop models were compared in their ability to estimate water needs and actual irrigation. The first crop model is based on air temperature as the main determining factor for the growth, whereas the second relies on high resolution data from Sentinel-2 satellite to monitor plant growth. Both are applied at plot scale in a watershed of 800 km2 characterized by irrigation withdrawals. Results show that including remote sensing data leads to more realistic modeled emergence dates for summer crops. However both approaches have proven to be able to reproduce the evolution of daily irrigation withdrawals throughout the year. As a result, both approaches allowed to simulate the downstream flow with a good daily accuracy, especially during low flow periods.
Fichier principal
Vignette du fichier
LGRS.2019.2960625.pdf (1.55 Mo) Télécharger le fichier
Origine Fichiers produits par l'(les) auteur(s)

Dates et versions

hal-04610578 , version 1 (26-06-2024)

Identifiants

Citer

Hassan Bazzi, Dino Ienco, Nicolas Baghdadi, Mehrez Zribi, Valérie Demarez. Distilling Before Refine: Spatio-Temporal Transfer Learning for Mapping Irrigated Areas Using Sentinel-1 Time Series. IEEE Geoscience and Remote Sensing Letters, 2020, 17 (11), pp.1909-1913. ⟨10.1109/LGRS.2019.2960625⟩. ⟨hal-04610578⟩
32 Consultations
0 Téléchargements

Altmetric

Partager

Gmail Mastodon Facebook X LinkedIn More