Circumpolar patterns of potential mean annual ground temperature based on surface state obtained from microwave satellite data
<p>Gap filling is required for temporally and spatially consistent records of land surface temperature from satellite data due to clouds or snow cover. Land surface state, frozen versus unfrozen conditions, can be, however, captured globally with satellite data obtained by microwave sensors...
| Main Authors: | , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
Copernicus Publications
2018-07-01
|
| Series: | The Cryosphere |
| Online Access: | https://www.the-cryosphere.net/12/2349/2018/tc-12-2349-2018.pdf |
| Summary: | <p>Gap filling is required for temporally and spatially consistent records of land
surface temperature from satellite data due to clouds or snow cover. Land
surface state, frozen versus unfrozen conditions, can be, however, captured
globally with satellite data obtained by microwave sensors. The number of
frozen days per year has been previously proposed to be used for permafrost
extent determination. This suggests an underlying relationship between number
of frozen days and mean annual ground temperature (MAGT). We tested this
hypothesis for the Northern Hemisphere north of 50° N using
coarse-spatial-resolution microwave satellite data (Metop Advanced
SCATterometer – ASCAT – and Special Sensor Microwave Imager – SSM/I; 12.5
and 25 km nominal resolution; 2007–2012), which provide the necessary
temporal sampling. The MAGT from GTN-P (Global Terrestrial Network for
Permafrost) borehole records at the coldest sensor depth was tested for
validity in order to build a comprehensive in situ data set for calibration
and validation and was eventually applied. Results are discussed with respect
to snow water equivalent, soil properties, land cover and permafrost type.
The obtained temperature maps were classified for permafrost extent and
compared to alternative approaches.</p><p>An <i>R</i><sup>2</sup> of 0.99 was found for correlation between and MAGT at zero annual
amplitude provided in GTN-P metadata and MAGT at the coldest sensor depth. The
latter could be obtained with an RMSE of 2.2 °C from ASCAT and
2.5 °C from SSM/I surface state records using a linear model. The
average deviation within the validation period is less than 1 °C at
locations without glaciers and coastlines within the resolution cell in the case
of ASCAT. The exclusion of snow melt days (available for ASCAT) led to better
results. This suggests that soil warming under wet snow cover needs to be
accounted for in this context. Specifically Scandinavia and western Russia
are affected. In addition, MAGT at the coldest sensor depth was overestimated in
areas with a certain amount of organic material and in areas of cold
permafrost. The derived permafrost extent differed between the used data sets
and methods. Deviations are high in central Siberia, for example. We show that
microwave-satellite-derived surface state records can provide an
estimation of not only permafrost extent but also MAGT without the need for gap
filling. This applies specifically to ASCAT. The deviations among the
tested data sets, their spatial patterns as well as in relation to
environmental conditions, revealed areas which need special attention for
modelling of MAGT.</p> |
|---|---|
| ISSN: | 1994-0416 1994-0424 |