Upscaling instantaneous to daily evapotranspiration using modelled daily shortwave radiation for remote sensing applications: an artificial neural network approach
Upscaling instantaneous evapotranspiration retrieved at any specific time-of-day (ET<sub><i>i</i></sub>) to daily evapotranspiration (ET<sub>d</sub>) is a key challenge in mapping regional ET using polar orbiting sensors. Various studies have unanimously cited the...
| Main Authors: | , , , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
Copernicus Publications
2017-01-01
|
| Series: | Hydrology and Earth System Sciences |
| Online Access: | http://www.hydrol-earth-syst-sci.net/21/197/2017/hess-21-197-2017.pdf |
| _version_ | 1811223288682119168 |
|---|---|
| author | L. Wandera K. Mallick G. Kiely O. Roupsard M. Peichl V. Magliulo |
| author_facet | L. Wandera K. Mallick G. Kiely O. Roupsard M. Peichl V. Magliulo |
| author_sort | L. Wandera |
| collection | DOAJ |
| description | Upscaling instantaneous evapotranspiration retrieved at any specific
time-of-day (ET<sub><i>i</i></sub>) to daily evapotranspiration
(ET<sub>d</sub>) is a key challenge in mapping regional ET using
polar orbiting sensors. Various studies have unanimously cited the shortwave incoming radiation (<i>R</i><sub>S</sub>) to be the most robust reference
variable explaining the ratio between ET<sub>d</sub> and
ET<sub><i>i</i></sub>. This study aims to contribute in ET<sub><i>i</i></sub>
upscaling for global studies using the ratio between daily and instantaneous
incoming shortwave radiation (<i>R</i><sub>Sd</sub> ∕ <i>R</i><sub>Si</sub>) as a
factor for converting ET<sub><i>i</i></sub> to ET<sub>d</sub>.<br><br>This paper proposes an artificial neural network (ANN) machine-learning
algorithm first to predict <i>R</i><sub>Sd</sub> from <i>R</i><sub>Si</sub> followed
by using the <i>R</i><sub>Sd</sub> ∕ <i>R</i><sub>Si</sub> ratio to convert
ET<sub><i>i</i></sub> to ET<sub>d</sub> across different terrestrial
ecosystems. Using <i>R</i><sub>Si</sub> and <i>R</i><sub>Sd</sub> observations from
multiple sub-networks of the FLUXNET database spread across different climates
and biomes (to represent inputs that would typically be obtainable from
remote sensors during the overpass time) in conjunction with some
astronomical variables (e.g. solar zenith angle, day length, exoatmospheric
shortwave radiation), we developed the ANN model for reproducing
<i>R</i><sub>Sd</sub> and further used it to upscale ET<sub><i>i</i></sub> to
ET<sub>d</sub>. The efficiency of the ANN is evaluated for different
morning and afternoon times of day, under varying sky conditions, and also at
different geographic locations. <i>R</i><sub>S</sub>-based upscaled
ET<sub>d</sub> produced a significant linear relation (<i>R</i><sup>2</sup> = 0.65
to 0.69), low bias (−0.31 to −0.56 MJ m<sup>−2</sup> d<sup>−1</sup>; approx. 4 %), and
good agreement (RMSE 1.55 to 1.86 MJ m<sup>−2</sup> d<sup>−1</sup>; approx. 10 %) with
the observed ET<sub>d</sub>, although a systematic overestimation of
ET<sub>d</sub> was also noted under persistent cloudy sky conditions.
Inclusion of soil moisture and rainfall information in ANN training reduced
the systematic overestimation tendency in predominantly overcast days. An
intercomparison with existing upscaling method at daily, 8-day, monthly, and
yearly temporal resolution revealed a robust performance of the ANN-driven
<i>R</i><sub>S</sub>-based ET<sub><i>i</i></sub> upscaling method and was found to
produce lowest RMSE under cloudy conditions. Sensitivity analysis revealed
variable sensitivity of the method to biome selection and high
ET<sub>d</sub> prediction errors in forest ecosystems are primarily
associated with greater rainfall and cloudiness. The overall methodology appears
to be promising and has substantial potential for upscaling
ET<sub><i>i</i></sub> to ET<sub>d</sub> for field and regional-scale
evapotranspiration mapping studies using polar orbiting satellites. |
| first_indexed | 2024-04-12T08:30:24Z |
| format | Article |
| id | doaj.art-0805e24e151b47a6a829e679efacd186 |
| institution | Directory Open Access Journal |
| issn | 1027-5606 1607-7938 |
| language | English |
| last_indexed | 2024-04-12T08:30:24Z |
| publishDate | 2017-01-01 |
| publisher | Copernicus Publications |
| record_format | Article |
| series | Hydrology and Earth System Sciences |
| spelling | doaj.art-0805e24e151b47a6a829e679efacd1862022-12-22T03:40:14ZengCopernicus PublicationsHydrology and Earth System Sciences1027-56061607-79382017-01-0121119721510.5194/hess-21-197-2017Upscaling instantaneous to daily evapotranspiration using modelled daily shortwave radiation for remote sensing applications: an artificial neural network approachL. Wandera0K. Mallick1G. Kiely2O. Roupsard3M. Peichl4V. Magliulo5Remote Sensing and Ecohydrological Modeling, Dept. ERIN, Luxembourg Institute of Science and Technology, Belvaux, LuxembourgRemote Sensing and Ecohydrological Modeling, Dept. ERIN, Luxembourg Institute of Science and Technology, Belvaux, LuxembourgCivil and Environmental Engineering Dept., and Environmental Research Institute, University College Cork, Cork, IrelandCIRAD, UMR Eco&Sols, 2 Place Viala, Montpellier, FranceDepartment of Forest Ecology and Management, Swedish University of Agricultural Sciences, Umeå, SwedenConsiglio Nazionale delle Ricerche, ISAFOM, Ercolano, Naples, ItalyUpscaling instantaneous evapotranspiration retrieved at any specific time-of-day (ET<sub><i>i</i></sub>) to daily evapotranspiration (ET<sub>d</sub>) is a key challenge in mapping regional ET using polar orbiting sensors. Various studies have unanimously cited the shortwave incoming radiation (<i>R</i><sub>S</sub>) to be the most robust reference variable explaining the ratio between ET<sub>d</sub> and ET<sub><i>i</i></sub>. This study aims to contribute in ET<sub><i>i</i></sub> upscaling for global studies using the ratio between daily and instantaneous incoming shortwave radiation (<i>R</i><sub>Sd</sub> ∕ <i>R</i><sub>Si</sub>) as a factor for converting ET<sub><i>i</i></sub> to ET<sub>d</sub>.<br><br>This paper proposes an artificial neural network (ANN) machine-learning algorithm first to predict <i>R</i><sub>Sd</sub> from <i>R</i><sub>Si</sub> followed by using the <i>R</i><sub>Sd</sub> ∕ <i>R</i><sub>Si</sub> ratio to convert ET<sub><i>i</i></sub> to ET<sub>d</sub> across different terrestrial ecosystems. Using <i>R</i><sub>Si</sub> and <i>R</i><sub>Sd</sub> observations from multiple sub-networks of the FLUXNET database spread across different climates and biomes (to represent inputs that would typically be obtainable from remote sensors during the overpass time) in conjunction with some astronomical variables (e.g. solar zenith angle, day length, exoatmospheric shortwave radiation), we developed the ANN model for reproducing <i>R</i><sub>Sd</sub> and further used it to upscale ET<sub><i>i</i></sub> to ET<sub>d</sub>. The efficiency of the ANN is evaluated for different morning and afternoon times of day, under varying sky conditions, and also at different geographic locations. <i>R</i><sub>S</sub>-based upscaled ET<sub>d</sub> produced a significant linear relation (<i>R</i><sup>2</sup> = 0.65 to 0.69), low bias (−0.31 to −0.56 MJ m<sup>−2</sup> d<sup>−1</sup>; approx. 4 %), and good agreement (RMSE 1.55 to 1.86 MJ m<sup>−2</sup> d<sup>−1</sup>; approx. 10 %) with the observed ET<sub>d</sub>, although a systematic overestimation of ET<sub>d</sub> was also noted under persistent cloudy sky conditions. Inclusion of soil moisture and rainfall information in ANN training reduced the systematic overestimation tendency in predominantly overcast days. An intercomparison with existing upscaling method at daily, 8-day, monthly, and yearly temporal resolution revealed a robust performance of the ANN-driven <i>R</i><sub>S</sub>-based ET<sub><i>i</i></sub> upscaling method and was found to produce lowest RMSE under cloudy conditions. Sensitivity analysis revealed variable sensitivity of the method to biome selection and high ET<sub>d</sub> prediction errors in forest ecosystems are primarily associated with greater rainfall and cloudiness. The overall methodology appears to be promising and has substantial potential for upscaling ET<sub><i>i</i></sub> to ET<sub>d</sub> for field and regional-scale evapotranspiration mapping studies using polar orbiting satellites.http://www.hydrol-earth-syst-sci.net/21/197/2017/hess-21-197-2017.pdf |
| spellingShingle | L. Wandera K. Mallick G. Kiely O. Roupsard M. Peichl V. Magliulo Upscaling instantaneous to daily evapotranspiration using modelled daily shortwave radiation for remote sensing applications: an artificial neural network approach Hydrology and Earth System Sciences |
| title | Upscaling instantaneous to daily evapotranspiration using modelled daily shortwave radiation for remote sensing applications: an artificial neural network approach |
| title_full | Upscaling instantaneous to daily evapotranspiration using modelled daily shortwave radiation for remote sensing applications: an artificial neural network approach |
| title_fullStr | Upscaling instantaneous to daily evapotranspiration using modelled daily shortwave radiation for remote sensing applications: an artificial neural network approach |
| title_full_unstemmed | Upscaling instantaneous to daily evapotranspiration using modelled daily shortwave radiation for remote sensing applications: an artificial neural network approach |
| title_short | Upscaling instantaneous to daily evapotranspiration using modelled daily shortwave radiation for remote sensing applications: an artificial neural network approach |
| title_sort | upscaling instantaneous to daily evapotranspiration using modelled daily shortwave radiation for remote sensing applications an artificial neural network approach |
| url | http://www.hydrol-earth-syst-sci.net/21/197/2017/hess-21-197-2017.pdf |
| work_keys_str_mv | AT lwandera upscalinginstantaneoustodailyevapotranspirationusingmodelleddailyshortwaveradiationforremotesensingapplicationsanartificialneuralnetworkapproach AT kmallick upscalinginstantaneoustodailyevapotranspirationusingmodelleddailyshortwaveradiationforremotesensingapplicationsanartificialneuralnetworkapproach AT gkiely upscalinginstantaneoustodailyevapotranspirationusingmodelleddailyshortwaveradiationforremotesensingapplicationsanartificialneuralnetworkapproach AT oroupsard upscalinginstantaneoustodailyevapotranspirationusingmodelleddailyshortwaveradiationforremotesensingapplicationsanartificialneuralnetworkapproach AT mpeichl upscalinginstantaneoustodailyevapotranspirationusingmodelleddailyshortwaveradiationforremotesensingapplicationsanartificialneuralnetworkapproach AT vmagliulo upscalinginstantaneoustodailyevapotranspirationusingmodelleddailyshortwaveradiationforremotesensingapplicationsanartificialneuralnetworkapproach |