Dehesa environment mapping with transference of a Random Forest classifier to neighboring ultra-high spatial resolution imagery at class and macro-class land cover levels

Campón, Laura Fragoso; Quirós, Elia; Gállego, José Antonio Gutiérrez

doi:10.1007/s00477-020-01880-3

Cited by 2 publications

(2 citation statements)

References 96 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…The land cover map used in this work (Figure 1) was developed in a previous study in which hydrological land cover categories were mapped based on the runoff generation capability of the land cover types using remote sensing techniques (Fragoso‐Campón, Quirós, & Gutiérrez Gallego, 2020). The land cover can be divided into three groups: forested, agricultural and impervious cover.…”

Section: Methodsmentioning

confidence: 99%

“…Forested land cover categories comprise evergreen forest, deciduous forest, dehesas , shrubs, and herbaceous vegetation. The dehesa is a typical cover from Extremadura (Fragoso‐Campón, Quirós, & Gutiérrez Gallego, 2020), defined by Devesa Alcaraz (1995) as ‘ pasturelands populated by holm and/or cork oaks, with an understorey of open grassland, cereal crops, or Mediterranean scrub ’. The agricultural land cover categories are rainfed crops (mainly olive trees, vineyards and cold‐season annual crops and cherry trees) and irrigated crops (seasonal warm crops).…”

Section: Methodsmentioning

confidence: 99%

See 1 more Smart Citation

Analysing the capability of a catchment's spectral signature to regionalize hydrological parameters

Campón

Durán‐Barroso

Quirós

2022

Hydrological Processes

Self Cite

View full text Add to dashboard Cite

Water resource management in ungauged catchments is complex due to uncertainties around the hydrological parameters that characterize streamflow behaviour. These parameters are usually defined by regionalization approaches, in which the hydrological response patterns of ungauged basins are inferred from those of gauged basins. Regression‐based methods using physical properties derived from cartographic data sources are widely used. The current remote sensing techniques offer new opportunities for the regionalization of hydrological parameters since the hydrological response depends on the physical attributes related to the spectral responses of a given land surface. Moreover, machine learning approaches have not been specifically applied to the regionalization of hydrological parameters in forested areas. This work studies the capability of a catchment's spectral signature based on Sentinel‐1 and Sentinel‐2 data to address a regression‐based regionalization of hydrological model parameters using a machine learning approach. Hydrological modelling was conducted by the HBV‐light model. We tested the random forest algorithm in several regionalization scenarios: the new approach using the catchments' spectral signature, the traditional method using physical properties and a fusion of these methods. The calibration results were excellent (median KGE = 0.83), and the regionalized parameters achieved good performance, in which the three scenarios showed almost the same goodness of fit (median KGE = 0.45–0.50). We found that the effectiveness depends on the climatic environment and that predictions in humid catchments exhibited better performance than those in the driest catchments. The physical approach (median KGE = 0.71) exhibited better performance than the spectral approach (median KGE = 0.64) in humid catchments, whereas spectral regionalization (median KGE = 0.33) enhanced the physical scenario in the driest catchments (median KGE = 0.25). Our results confirm that regionalization is still challenging in drier climates, such as in the Mediterranean environment. The new spectral approach showed promising results and it was effective in the analysis of the relationship between the spectral response of the territory and its hydrological characteristics, specially, where no cartographic data is available.

show abstract

Section: Methodsmentioning

confidence: 99%

Section: Methodsmentioning

confidence: 99%