Content of Figures .............................................. 5
Content of Tables ............................................... 8
List of Shortcuts and Abbreviations ............................. 9
1. Forecasting of earthquakes - An introduction ................ 11
2. Present status in prediction of earthquakes by remote
sensing ..................................................... 12
2.1. Geo-seismic echo on solar activity ..................... 13
2.2. Ionospheric processes .................................. 14
3. Analysis of exemplarily earthquake precursor events using
space-borne infrared and visible remote sensing data sets ... 16
3.1. Characteristics, possible nature and detection
problems of thermal anomalies related to earthquake
preparation process .................................... 16
3.2. Review of the algorithms used to extract pre-seismic
thermal anomalies ...................................... 18
3.3. Testing the multi-temporal approach for monitoring
thermal anomalies ...................................... 19
3.3.1. Anomalies in sea surface temperature ............ 19
3.3.1.1. The case of the Kocaeli (Izmit)
earthquake of August, 17, 1999 ......... 19
3.3.1.2. The case of the Greek earthquake
of August, 14, 2003 .................... 25
3.3.2. Anomalies in night-time land surface
temperature ..................................... 26
3.3.3. Anomalies in daytime land surface temperature ... 35
3.4. Wavelet approaches ..................................... 36
3.4.1. The "CQuake" wavelet algorithm .................. 36
3.4.2. The two-dimensional multi-resolution wavelet
analysis ........................................ 37
3.4.2.1. Sea surface temperature anomalies
at Kocaeli (Izmit) in August 1999 ...... 37
3.4.2.2. Sea surface temperature anomalies
at Ionian Sea in August 2003 ........... 39
3.5. Characteristics of visible phenomena ................... 40
3.6. Conclusions on used methods and approaches ............. 41
3.6.1. Monitoring thermal anomalies by RETIRA multi-
temporal approaches ............................. 41
3.6.2. Monitoring thermal anomalies by two-
dimensional multi-resolution wavelet analysis ... 43
4. Requirements to dedicated micro-satellite based IR sensor
systems for earthquake precursor detection .................. 44
4.1. Spectro-radiometric requirements on IR sensors for
space-borne detection of thermal precursors of
earthquakes ............................................ 44
4.2. Spectro-radiometric requirements to space-borne
sensors for detection of linear cloud structures
occurring before earthquakes ........................... 45
4.3. Prospective of an Imaging Thermal Infra-Red (TIR)
spectrometry for detection of thermal anomalies in
arid and semi-arid terrain ............................. 45
4.4. Overall mission requirement aspects .................... 46
5. Conclusions and recommendations ............................. 48
Annexes ........................................................ 50
Annex A. Evaluation of the error sources in the multi-
temporal approach for monitoring thermal anomalies
in seismically active areas ........................... 51
A.1. Atmospheric effects ................................... 51
A.2. Surface emissivity .................................... 54
A.3. Temporal and spatial surface temperature
variability ........................................... 57
A.4. Geo-referencing errors ................................ 63
Annex B. Development of approaches to minimize the effect
of error sources in the multi-temporal approach
for monitoring thermal anomalies in seismically
active areas .......................................... 65
B.1. Atmospheric correction ................................ 65
B.2. Accounting for surface emissivity ..................... 70
B.3. Reducing the effect of surface temporal variability ... 73
B.4. Recommendations on method utilization and sensor
observation features .................................. 73
Annex С. Development of methods for reliable identification
of typical cloud structures from space as potential
earthquake precursors ................................. 75
C.1. Review of cloud detection algorithms .................. 75
C.2. Simulation of multi-spectral radiometric
characteristics of clouds ............................. 77
C.3. Cloud detection algorithm ............................. 80
C.4. Aspects of linear cloud structure recognition ......... 83
C.5. Sensor requirements for cloud detection ............... 83
Annex D. Development of proposals for dedicated IR sensor
systems for earthquake precursor detection by
micro-satellites ...................................... 85
D.1. Scope ................................................. 85
D.2. Currently used IR sensors for earthquake precursor
detection ............................................. 85
D.3. Upcoming European IR sensors with the potential
for earthquake precursor detection .................... 86
D.4. Prospective German micro-satellite missions with
IR sensors ............................................ 87
D.5. Concept proposals for a dedicated IR sensor system
for earthquake precursor detection by micro-
satellites ............................................ 88
D.5.1. Concept A: Sensor supplement to
a prospective fire monitoring constellation -
a narrow cost proposal ......................... 88
D.5.2. Concept B: thermal IR imaging spectrometer ..... 92
References ..................................................... 98
Content of Figures
Fig. 2.1: Seismo-atmospheric-ionospheric couplings according
to Pulinets (Pulinets, 2004) ........................ 15
Fig. 3.1: Sea surface temperature distribution over the
Aegean Sea. The selected anomaly area is confined
with blue ............................................ 20
Fig. 3.2: SST RETIRA index distribution over the Aegean Sea.
The selected anomaly area is confined with blue ...... 21
Fig. 3.3: The total area of SST RETIRA anomalies with
RETIRA > 2.5 in the Aegean Sea as a function of
time: a - for the entire available observation
period; b - in August, 1999 (the month of the
earthquake). The dotted line indicates August,
17, 1999 (the earthquake date). The arrow
corresponds to the RETIRA anomalies on August,
13, 1999 ............................................. 22
Fig. 3.4: The total area of SST RETIRA anomalies with
RETIRA > 2.5 in the Aegean Sea in the month of
August during 8 years of observations. The dotted
line indicates the earthquake date (August, 17,
1999). The arrow corresponds to the RETIRA
anomalies on August, 13, 1999 ........................ 23
Fig. 3.5: Time series of SST (a, b) and of RETIRA (c, d) in
the area selected in Fig. 3.2 and Fig.3.3. The
dotted line indicates the earthquake date (August,
17, 1999). The arrow corresponds to the date of
the RETIRA anomaly of August, 13, 1999 ............... 24
Fig. 3.6: Map of the region of the earthquake. The
epicenters of the earthquakes on August, 14, 2003
and March, 1, 2004 are marked with a star.
The plate boundary, major fault lines and the
type of faulting are also indicated (Cervone etal,
2004) ................................................ 25
Fig. 3.7: Sea surface temperature (a) and RETIRA index (b)
distribution in the Ionian Sea on August, 6, 2003.
Two selected areas are confined with blue ............ 27
Fig. 3.8: Sea surface temperature (a) and RETIRA index (b)
distribution in the Ionian Sea on August, 13, 2003 ... 28
Fig. 3.9: Time series of SST (a, b) and of RETIRA (c, d)
values in area 1 in Fig.3.6. The dotted line
indicates the earthquake date (August, 14, 2003).
The arrow corresponds to the date of the RETIRA
anomaly of August, 6, 2003 ........................... 29
Fig.3.10: Time series of SST (a, b) and of RETIRA (c, d) in
area 2 in Fig.3.6. The dotted line indicates
the earthquake date (August, 14, 2003). The arrow
corresponds to the date of the RETIRA anomaly of
August, 6, 2003 ...................................... 30
Fig.3.11: Correlation of SST RETIRA peaks in areas 1 and 2
in Fig.3.6: a - RETIRA time series in area 1;
b - RETIRA time series in area 2; с - summary value
of the RETIRA peaks with magnitude > 2 in both
areas; d - summary value of the RETIRA peaks with
magnitude > 3 in both areas .......................... 31
Fig.3.12: Night-time LST-based RETIRA index distribution
over Greece on August, 12-15, 2003 ................... 32
Fig.3.13: The total area of night-time LST RETIRA anomalies
with RETIRA > 2 over Greece as a function of
time: a - for the entire available observation
period; b - for the month of the earthquake
(August 2003). The dotted line indicates the
earthquake date. The arrow corresponds to the
RETIRA anomalies on August, 13, 2003 ................. 33
Fig.3.14: The total area of night-time LST-based RETIRA
anomalies with RETIRA > 2 over Greece in the month
of August during 8 years of observations. The
dotted line indicates the earthquake date
(14 August 2003). The arrow corresponds to the
RETIRA anomalies on 13 August 1999 ................... 34
Fig.3.15: The total area of daytime LST-based RETIRA
anomalies with RETIRA > 2 over Greece as a function
of time: a - for the entire available observation
period; b - for the year of the earthquake (2003);
с - for the month of the earthquake (August 2003).
The dotted line indicates the earthquake date
(14 August 2003) ..................................... 35
Fig.3.16: Multi-resolution analysis by Daubechies wavelet
2 of SST of Aegean Sea at August 13, 1999. The
arrow indicates a possible thermal anomaly.
Calculations are performed with (ITT, 2007) .......... 38
Fig.3.17: Low pass SST images at scale 2 at (a) August 10,
and (b) August 13, 1999 of the Aegean Sea. White
marked areas indicate occurring thermal anomalies .... 39
Fig.3.18: Low pass SST images at scale 2 at (a) August 06,
and (b) August 13, 2003 of the Ionian Sea. White
marked areas indicate occurring thermal anomalies .... 39
Fig.3.19: Linear cloud structure (indicated with the yellow
arrow) in the SST distribution south of the Ionian
Sea on February, 28, 2004 ............................ 40
Fig.3.20: Multi-resolution analysis by Haar wavelet of SST
scene in the Ionian Sea from February 29, 2004.
Calculations are performed with (ITT, 2007) .......... 42
Fig.3.21: Wavelet power spectrum of SST scene in the Ionian
Sea from February 29, 2004. The arrow in the high
pass sub-band at scale 2 indicates the linear cloud
structure. Calculations are performed with (ITT,
2007) ................................................ 43
Fig. А.1: Transmittance of atmospheric constituents and
the total atmospheric transmittance in the thermal
infra-red range (mid-latitude summer model with
rural aerosol, vis = 23 km) .......................... 51
Fig. A.2: The atmospheric effect on NOAA-7 AVHRR band
temperatures in different models (Wan, 1999):
T4 and T5 is at-sensor brightness temperature in
AVHRR channels 4(11 pm) and 5 (12pm), Ts is the
surface temperature .................................. 52
Fig. A.3: Atmospheric transmission (a), path radiance (b)
and at-sensor brightness temperature (c) of a
surface with at-ground brightness temperature of
300 К as functions of the off-nadir viewing angle
(mid-latitude summer model with rural aerosol,
vis = 23 km) ......................................... 53
Fig. A.4: At-ground and at-sensor brightness temperature of
a surface with a thermodynamic temperature of
300 К as a function of surface emissivity (mid-
latitude summer atmospheric model with a rural
aerosol, vis = 23 km) ................................ 54
Fig. A.5: Spectral emissivity of typical natural objects
from the ASTER spectral library in the 8-14pm
spectral range ....................................... 55
Fig. A.6: Emissivity of objects from the ASTER spectral
library in the MODIS 11 pm spectral channel and
the emissivity difference in the MODIS 11 pm and
12 pm spectral channels .............................. 56
Fig. A.7: Emissivity of water in the MODIS 11 pm spectral
band as a function of the incidence angle ............ 57
Fig. A.8: Seasonal variation of the mean regional
temperature: a - sea surface temperature over
the Aegean See; b - daytime land surface
temperature over Greece; с - night-time land
surface temperature over Greece ...................... 59
Fig. A.9: Diurnal variation of the land surface temperature.
(http://apollo.lsc.vsc.edu/classes/met130/notes/
chapter3/daily trend5.html) .......................... 60
Fig.A.10: Day-night-time difference of the mean regional
temperature over Greece .............................. 60
Fig.A.11: Standard deviation (SD) of the spatial surface
temperature variation: a - sea surface temperature
over the Aegean See; b - daytime land surface
temperature over Greece; с - night time land
surface temperature over Greece ...................... 61
Fig.A.12: Standard deviation of the "meteorological
variations" of the surface temperature: a -sea
surface temperature over the Aegean See; b -
daytime land surface temperature over Greece;
с - night-time land surface temperature over
Greece; solid line - total effect, dashed line -
after subtraction of the regional effect ............. 62
Fig.A.13: Standard deviation of the surface temperature
after a shift of 1 pixel along a diagonal: a -
sea surface temperature over the Aegean See;
b - daytime land surface temperature over Greece;
с - night-time land surface temperature over
Greece ............................................... 64
Fig. B.1: An empirical relation between tmm and MMD based
on 86 laboratory reflectance spectra of rocks,
soils, vegetation, snow and water (Gillespie et
al, 1999) ............................................ 72
Fig. С.1: Spectral characteristics of standard MODTRAN clouds
and fogs, vegetation and snow (a) in the VIS-SWIR
and (b) in the MIR-TIR spectral ranges ............... 78
Fig. C.2: Clouds and other objects in a diagram showing TIR
temperature & VIS+NIR reflectance. The cloud
detection region according to the IGPB algorithm
is blue shaded ....................................... 79
Fig. C.3: Clouds and other objects in diagram of TIR
temperature & MIR reflectance ........................ 79
Fig. C.4: Cloud detection in a MODIS image: a - VIS channel 1
(0.62-0.67pm); b - TIR channel 31 (10.78-
11.28 pm) ............................................ 81
Fig. C.5: Map of the region of the earthquake. The epicenters
of the earthquakes on August, 14, 2003 and March,
1, 2004 are marked with a star. The plate boundary,
major fault lines and the type of faulting are also
indicated (Cervone et al, 2004) ...................... 84
Fig. С.6: Linear cloud structure (indicated with the yellow
arrow) in the SST distribution south of the Ionian
Sea on 28 February 2004 .............................. 84
Fig. D.1: Optical scheme of MIBS ............................... 94
Fig. D.2: Overall view of the MIBS bread board (Leijtens &
de Goeij, 2005) ...................................... 94
Fig. D.3: The main all-reflective optics elements of MERTIS,
consisting of the Three Mirror Anastigmat (TMA)
and the Offner spectrometer together with the
micro-bolometer 2-D array detector focal plane ....... 95
Fig. D.4: Overall view of MERTIS (Walter et al, 2006) .......... 96
Content of Tables
Table 3.1: Main natural and observational factors affecting
TIR signal (Tramutoli et al, 2005) .................. 17
Table B.1: NOAA-18 regression coefficients for different
equations (http://noaasis.noaa.gov/NOAASIS/pubs/
SST/noaa18sst.txt) .................................. 67
Table B.2: Error estimate of the generalized split-window
MODIS LST algorithm (Wan and Dozier, 1996; Wan,
1999) ............................................... 69
Table D.1: Several observation characteristics of sensors
currently used for detection of thermal anomalies
as Earthquake precursors ............................ 86
Table D.2: Assumed spatial sampling intervals and swath
widths of the FMC main sensor and the
supplementary two band imaging TIR sensors .......... 90
Table D.3: GMES IR Element TIR instrument main
characteristics - from (GMES Infrared Element,
2005) ............................................... 91
|
