MODIS Land Surface Temperatures (LST)
From Land Surface Temperature (LST) maps, derivable indicators include temperature minima and maxima at annual, monthly, and decadal periods, the identification of late frost periods, unusual hot summers, growing degree days, spring temperature increase and autumnal temperature decrease (Neteler, 2005; Rizzoli et al., 2007; Carpi et al., 2008; Neteler, 2010).
While the raw data are of limited interest to landscape epidemiological applications, time series aggregation of the new sensor data leads to a new quality of ecological indicators which have not been available earlier. A special challenge is the complex terrain as it dominates the Southern Alps in Italy. It requires special attention to data processing and outlier detection.
MODIS LST (7 Apr 2006, 13:30) satellite with QA map and outlier detection applied:
cirrus cloud fields remain undetected
MODIS LST (7 Apr 2006, 13:30) satellite with second outlier detection applied:
cirrus cloud fields removed
MODIS LST (7 Apr 2006, 13:30) satellite with reinterpolated with
elevation as additional variable and exposition correction (preliminary results)
Temperature time series from MODIS LST
We use above method to produce reconstructed time series from MODIS LST. Here some preliminary results:
Raw (blue) and reconstructed (red) MODIS LST time at Arco (TN), Italy (click to enlarge)
To better understand the quality of the reconstruction, see below a close-by meteorological station. The reconstruction above is completely independent from the meteorological data and only based on remote sensing data.
Tmin at 2m from meteorological station at Arco (TN), Italy (click to enlarge)
The developed method will be published in detail in 2012 after acceptance of some papers.
European daily LST maps
We are currently working on the reconstruction of all European MODIS LST maps (~ 13,000 maps):
Reconstructed MODIS Land Surface Temperature (LST) map,
2003-07-09 at 10:30 (Terra satellite overpass) - click to enlarge
Processing the LST data
The European LST data reconstruction is very intensive since a series of input maps are used to produce the resulting reconstructed map (415 million pixels per map). We use our high performance computing GIS cluster for this and enjoy the flexibility of open source.
GIS Cluster (300 nodes) of PGIS @ FEM
LST Details: QC maps
The filtering of MODIS LST maps with the QC map bitpatterns is a bit tricky. Here a list of relevant bit combinations (table inspired by this course with corrections):
| Bit position | 7 | 6 | 5 | 4 | 3 | 2 | 1 | 0 | Notes |
| Integer | 128 | 64 | 32 | 16 | 8 | 4 | 2 | 1 | |
| 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | Good Quality, average LST <=1k |
| 1 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 1 | LST produced, other quality, recommend examination of more detailed QA |
| 2 | 0 | 0 | 0 | 0 | 0 | 0 | 1 | 0 | Not Produced, due to cloud effects |
| 3 | 0 | 0 | 0 | 0 | 0 | 0 | 1 | 1 | Not Produced, primarily due to reasons other than cloud |
| 5 | 0 | 0 | 0 | 0 | 0 | 1 | 0 | 1 | LST produced, other quality, recommend examination of more detailed QA |
| 17 | 0 | 0 | 0 | 1 | 0 | 0 | 0 | 1 | LST produced, other quality, Average emissivity error <= 0.02 |
| 21 | 0 | 0 | 0 | 1 | 0 | 1 | 0 | 1 | LST produced, other quality, Average emissivity error <= 0.02 |
| 64 | 0 | 1 | 0 | 0 | 0 | 0 | 0 | 0 | Good Quality, average LST <=2k |
| 65 | 0 | 1 | 0 | 0 | 0 | 0 | 0 | 1 | LST produced, other quality, Average emissivity error <= 0.01, Average LST error <= 2K |
| 69 | 0 | 1 | 0 | 0 | 0 | 1 | 0 | 1 | LST produced, other quality, Average emissivity error <= 0.01, Average LST error <= 2K |
| 81 | 0 | 1 | 0 | 1 | 0 | 0 | 0 | 1 | LST produced, other quality, Average emissivity error <= 0.02, Average LST error <= 2K |
| 85 | 0 | 1 | 0 | 1 | 0 | 1 | 0 | 1 | LST produced, other quality, Average emissivity error <= 0.02, Average LST error <= 2K |
| 129 | 1 | 0 | 0 | 0 | 0 | 0 | 0 | 1 | LST produced, other quality, Average emissivity error <= 0.01, Average LST error <= 3K |
| 133 | 1 | 0 | 0 | 0 | 0 | 1 | 0 | 1 | LST produced, other quality, Average emissivity error <= 0.01, Average LST error <= 3K |
| 145 | 1 | 0 | 0 | 1 | 0 | 0 | 0 | 1 | LST produced, other quality, Average emissivity error <= 0.02, Average LST error <= 3K |
| 149 | 1 | 0 | 0 | 1 | 0 | 1 | 0 | 1 | LST produced, other quality, Average emissivity error <= 0.02, Average LST error <= 3K |
The color indications are recommendations for the QC pixel values (red: LST pixel to be rejected; orange: LST pixel probably acceptable; green: LST pixel ok). The decision to accept or reject orange indicated values depends on you. Find here a decimal to binary converter. The bitpattern 65 appears to be hard to decide: in some maps it depicts poor pixel quality, in many others acceptable pixel quality. Hence we suggest our MODIS LST reconstruction algorithm.
Related publications
- °Neteler, M., °Roiz, D., Rocchini, D., Castellani, C. and Rizzoli, A. (2011). Terra and Aqua satellites track tiger mosquito invasion: modeling the potential distribution of Aedes albopictus in north-eastern Italy. International Journal of Health Geographics, 10:49 (in press) [ Abstract | DOI | PDF ] (IF: 2.34)
°The authors contributed equally - °Roiz D., °Neteler M., Castellani C., Arnoldi D., Rizzoli A. (2011). Climatic Factors Driving Invasion of the Tiger Mosquito (Aedes albopictus) into New Areas of Trentino, Northern Italy. PLoS ONE. 6(4): e14800. [DOI | PDF] (IF: 4.411) - press reactions
°The authors contributed equally - Neteler, M., 2010: Estimating daily Land Surface Temperatures in mountainous environments by reconstructed MODIS LST data. Remote Sensing 2(1), 333-351. [ DOI | Abstract | PDF ]
- Neteler, M., 2010: Spatio-temporal reconstruction of satellite-based temperature maps and their application to the prediction of tick and mosquito disease vector distribution in Northern Italy. EDEN PhD summary. Catalogued by the EDEN EU/FP6 Steering Committee as EDEN0176
- Carpi G., Cagnacci F., Neteler M., Rizzoli A, 2008: Tick infestation on roe deer in relation to geographic and remotely-sensed climatic variables in a tick-borne encephalitis endemic area. Epidemiology and Infection,136, pp. 1416-1424. (DOI) (ISI 2008: 2.36) [ PubMed ]
- A. Rizzoli, M. Neteler, R. Rosà, W. Versini, A. Cristofolini, M. Bregoli, A. Buckley, and E.A. Gould, 2007: Early detection of TBEv spatial distribution and activity in the Province of Trento assessed using serological and remotely-sensed climatic data. Geospatial Health, 1(2), pp. 169-176. [ PubMed | PDF ]
- M. Neteler, 2005: Time series processing of MODIS satellite data for landscape epidemiological applications. International Journal of Geoinformatics, 1(1), pp. 133-138 (PDF)
Previous page: MODIS NDVI/EVI analysis
Next page: MODIS snow
