Multi-sensor satellite observations in support of Arctic Bird Habitat Characterization. Valentijn Venus, Andrew Skidmore, Bert Toxopeus Natural Resources Department, ITC. Remote sensing over Arctic's. Hostile environment, except for some birds (and satellites)
Multi-sensor satellite observations in support ofArctic Bird Habitat Characterization
Valentijn Venus, Andrew Skidmore, Bert Toxopeus
Natural Resources Department, ITC
21:50 (drift -3.2 min/month)
21:33 (drift -2.4 min/month)
21:31 (drift -0.1 min/month)
21:25 (drift -0.2 min/month)
17:33 (drift +4.8 min/month)
16:50 (drift -1.0 min/month)
13:57 (drift -1.6 min/month)
13:42 (drift 0.7 min/month)
13:23 (drift 0.3 min/month)
Mean Local Times at the Ascending Node (hh:mm)
Constellation as of 02 April 2009
pairs of POES satellites pass their orbital intersections within a few seconds in the polar regions
Occurs regularly in the +/- 70 to 80 latitude
AVHRR/N18 MODIS/Aqua Sample area
Reflectance Min Max Mean Stdev
Band 1 AVHRR 0.4301 0.4728 0.4523 0.008894
Band 1 MODIS 0.4800 0.5401 0.5113 0.012135
For this area with 205 samples, the difference between MODIS and AVHRR is about 13%, at 99% confidence level with uncertainty +/-0.4%. Spectral differences is not the main contributor to this discrepancy, according to radiative transfer calculations. Good example of calibration traceability issue.
SNO VIS/NIR example
Lat=79.82, SZA=82.339996, cos(sza)=0.13, TimeDiff 26 sec, Uncertainty due to SZA diff 0.1%,
Coverage every 5 minutes!
Meteosat First Generation
Rapid Scanning Service (RS) (10° E)
IODC Backup (67.5° E)
Primary Service (0° E)
IODC (57.5° E)
Three angles affect the signal received by a geo-stationary satellite sensor: 1) the solar zenith angle θ, 2) the satellite zenith angle Φ, and 3) the ‘co-scattering angle’ Ψ, between the direction towards the satellite and the sun as seen from ground. This information, which is unique for every ‘pixel’ and 5-minute satellite image, is used to correct to correct the signal for enhanced product generations (i.e. satellite estimated solar radiation).
large warm water
small cold ice
small cold water
large cold ice
MODIS surface temperature
(1 km resolution)
SEVIRI cloud types
(3 km resolution)
Instantaneous surface temperature derived from SEVIRI observations with the an enhanced four-channel algorithm (upper) and daily composite surface temperature derived from MODIS observations on 09/27/2004 (lower), over Europe.
Scatter plot of LST derived from the SEVIRI compared with that from the MODIS observations.
Scatter plot of LST derived from a ‘new’ 4-channel SEVIRI algorithm compared with that from ground observations.
The optimal model predicted observed budburst very accurately: r2 = 0.92 and 0.90, and root mean square error = 6.9 days and 7.5 days for a calibration and a validation dataset, respectively. Results predict that the average budburst in northern Europe in 2080-2099 will be 20 days (standard deviation (S.D.) = 3 days), 18 days (S.D. = 4 days) or 12 days (S.D. = 4 days) earlier than in the period 1980-1999, for 3 different climate change scenarios respectively.
Summarized phenology of the Fopin biosphere reserve as detected by MODIS NDVI after processing with TIMESAT. RPD (relative phenological development) is a rescaled version of the NDVI (see main text for details). The solid white lines shows the average altitudinal movement of 6 radiotracked giant pandas, thin white lines indicate the standard error of the average. Time slices of RPD during spring (1), summer (2) and autumn (3) with the positions of the radiotracked giant pandas during a 10 day period indicated by black markers.
+/- 35K EUR: HP ProLiant DL785 G6 Server