Downscaled land surface temperatures obtained by a particle smoothing approach, with links to NetCDF files


Autoria(s): André, Cyrille; Ottlé, Catherine; Mechri, Rihab; Pannekoucke, Olivier; Maignan, Fabienne
Cobertura

MEDIAN LATITUDE: 69.313365 * MEDIAN LONGITUDE: 145.398696 * SOUTH-BOUND LATITUDE: 62.305000 * WEST-BOUND LONGITUDE: 18.010000 * NORTH-BOUND LATITUDE: 78.900000 * EAST-BOUND LONGITUDE: -20.552500 * DATE/TIME START: 2001-07-15T00:00:00 * DATE/TIME END: 2009-09-15T00:00:00 * MINIMUM ELEVATION: 8.0 m * MAXIMUM ELEVATION: 1206.0 m

Data(s)

23/02/2015

Resumo

This dataset contains continuous time series of land surface temperature (LST) at spatial resolution of 300m around the 12 experimental sites of the PAGE21 project (grant agreement number 282700, funded by the EC seventh Framework Program theme FP7-ENV-2011). This dataset was produced from hourly LST time series at 25km scale, retrieved from SSM/I data (André et al., 2015, doi:10.1016/j.rse.2015.01.028) and downscaled to 300m using a dynamic model and a particle smoothing approach. This methodology is based on two main assumptions. First, LST spatial variability is mostly explained by land cover and soil hydric state. Second, LST is unique for a land cover class within the low resolution pixel. Given these hypotheses, this variable can be estimated using a land cover map and a physically based land surface model constrained with observations using a data assimilation process. This methodology described in Mechri et al. (2014, doi:10.1002/2013JD020354) was applied to the ORCHIDEE land surface model (Krinner et al., 2005, doi:10.1029/2003GB002199) to estimate prior values of each land cover class provided by the ESA CCI-Land Cover product (Bontemps et al., 2013) at 300m resolution . The assimilation process (particle smoother) consists in simulating ensemble of LST time series for each land cover class and for a large number of parameter sets. For each parameter set, the resulting temperatures are aggregated considering the grid fraction of each land cover and compared to the coarse observations. Miniminizing the distance between the aggregated model solutions and the observations allow us to select the simulated LST and the corresponding parameter sets which fit the observations most closely. The retained parameter sets are then duplicated and randomly perturbed before simulating the next time window. At the end, the most likely LST of each land cover class are estimated and used to reconstruct LST maps at 300m resolution using ESA CCI-Land Cover. The resulting temperature maps on which ice pixels were masked, are provided at daily time step during the nine-year analysis period (2000-2009).

Formato

text/tab-separated-values, 972 data points

Identificador

https://doi.pangaea.de/10.1594/PANGAEA.843151

doi:10.1594/PANGAEA.843151

Idioma(s)

en

Publicador

PANGAEA

Direitos

CC-BY: Creative Commons Attribution 3.0 Unported

Access constraints: unrestricted

Fonte

Supplement to: André, Cyrille; Ottlé, Catherine; Mechri, Rihab; Pannekoucke, Olivier; Maignan, Fabienne (2015): SSM/I Land Surface Temperature downscaling over Arctic sites. in preparation

Palavras-Chave #ABI; Abisko; Abisko, Lappland, northern Sweden; Alaska, USA; Changing Permafrost in the Arctic and its Global Effects in the 21st Century; CHE; Cherskii; DAR; Daring_lake; DATE/TIME; Event label; File name; File size; HER; Herschel_Island; KYT; Kytalyk; Kytalyk, Indigirka lowlands, Siberia; Latitude of event; Longitude of event; MULT; Multiple investigations; Nadym; North_Slope; PAGE21; Samoylov_Island; Samoylov Island, Lena Delta, Siberia; SPA; Spasskaya_Pad; Svalbard; Uniform resource locator/link to file; VOR; Vorkuta_Saida; Yamal Peninsula, northwestern Siberia; ZAC; Zackenberg; Zackenberg, Northeast Greenland
Tipo

Dataset