ORI GIN AL PA PER

Earthquake prediction activities and Damavandearthquake precursor test site in Iran

Mohammad Mokhtari

Received: 13 September 2008 / Accepted: 27 February 2009 / Published online: 27 March 2009� Springer Science+Business Media B.V. 2009

Abstract Iran has long been known as one of the most seismically active areas of the

world, and it frequently suffers destructive and catastrophic earthquakes that cause heavy

loss of human life and widespread damage. The Alborz region in the northern part of Iran

is an active EW trending mountain belt of 100 km wide and 600 km long. The Alborz

range is bounded by the Talesh Mountains to the west and the Kopet Dagh Mountains to

the east and consists of several sedimentary and volcanic layers of Cambrian to Eocene

ages that were deformed during the late Cenozoic collision. Several active faults affect the

central Alborz. The main active faults are the North Tehran and Mosha faults. The Mosha

fault is one of the major active faults in the central Alborz as shown by its strong historical

seismicity and its clear morphological signature. Situated in the vicinity of Tehran city, this

150-km-long N100� E trending fault represents an important potential seismic source. For

earthquake monitoring and possible future prediction/precursory purposes, a test site has

been established in the Alborz mountain region. The proximity to the capital of Iran with

its high population density, low frequency but high magnitude earthquake occurrence, and

active faults with their historical earthquake events have been considered as the main

criteria for this selection. In addition, within the test site, there are hot springs and deep

water wells that can be used for physico-chemical and radon gas analysis for earthquake

precursory studies. The present activities include magnetic measurements; application of

methodology for identification of seismogenic nodes for earthquakes of M C 6.0 in

the Alborz region developed by International Institute of Earthquake Prediction Theory

and Mathematical Geophysics, IIEPT RAS, Russian Academy of Science, Moscow

(IIEPT&MG RAS); a feasibility study using a dense seismic network for identification of

future locations of seismic monitoring stations and application of short-term prediction of

medium- and large-size earthquakes is based on Markov and extended self-similarity

analysis of seismic data. The establishment of the test site is ongoing, and the methodology

has been selected based on the IASPEI evaluation report on the most important precursors

with installation of (i) a local dense seismic network consisting of 25 short-period seis-

mometers, (ii) a GPS network consisting of eight instruments with 70 stations, (iii)

M. Mokhtari (&)International Institute of Earthquake Engineering and Seismology (IIEES), Tehran, Irane-mail: mokhtari@iiees.ac.ir

123

Nat Hazards (2010) 52:351–368DOI 10.1007/s11069-009-9375-2

magnetic network with four instruments, and (iv) radon gas and a physico-chemical study

on the springs and deep water wells.

Keywords Earthquake prediction � Test site � Precursors monitoring �Mosha fault � Seismogenic nodes

1 Introduction

Earthquakes are feared because they strike suddenly. Yet, there are reports of many kinds

of pre-earthquake signals. Widespread disagreement exists how these signals can be

generated in the Earth’s crust and whether they are in fact precursory or related to the

build-up of stress before major seismic events. Progress in understanding these signals has

been hindered by the lack of the underlying physical process or processes (Freund 2007a).

Over the course of more than a century, seismologists have learned how to use earth-

quakes as tool for revealing the interior of the Earth. Information extracted from the

propagation of seismic waves through the Earth has produced valuable insights into the

hidden structure of our dynamic planet. Unfortunately, earthquakes occur at unpredicted

times and in unpredicted places. Therefore, seismologists have made major efforts to find

ways to predict—within limits as narrow as possible—time, place, and magnitude of major

seismic events (Gokhberg et al. 1995; Lomnitz 1994; Milne 1899; Sykes, et al. 1999;

Turcotte 1991; Wyss and Dmowska 1997).

Seismological models have become ever more sophisticated in recent years with

some successful results (Holliday et al. 2005; Keilis-Borok 2002; Keilis-Borok and

Soloviev 2003; Rundle et al. 2003; Shebalin et al. 2004), but this tool alone has not

been able to recognize the microscopic processes that accompany the accumulation of

stresses before a rupture occurs. At the same time, it has been known for decades,

centuries, and even millennia that the earth sends out an incomprehensible array of

non-seismic signals before major seismic events (Tributsch 1984). Recording these

signals, understanding how they are generated, and what type of information they may

provide about impending earthquake activity have remained an indefinable goal (Ber-

nard et al. 1997; Hough 2002; Kagan 1997; Kanamori 1996; Knopoff 1996; Masood

1995; Park 1997; Uyeda 1998).

After the devastating Bam earthquake in Iran, the National Center for Earthquake

Prediction (NCEP) was established by government decree at the International Institute of

Earthquake Engineering and Seismology (IIEES). As part of NCEP, a test site has been

established in the Alborz mountain region for earthquake monitoring and possible future

prediction/precursor purposes. The proximity to the capital of Iran with its high population

density, low frequency but high magnitude earthquake occurrence, and active faults with

historical earthquake events are the main criteria for this selection. In addition, within the

test site, the existence of hot springs and deep water wells can be used for physico-

chemical and radon gas analysis for earthquake precursory studies.

In this article, we review the tectonic setting of the earthquake precursory test site in

the central Alborz. We discuss the possible use of earthquake precursors (Wyss 1991)

and the main methodology based on the International Association for Seismology and

Physics of Earth’s Interior (IASPEI) sub-commission’s recommendations. These

methodologies include monitoring of seismic, GPS, magnetic, and well or spring water

changes.

352 Nat Hazards (2010) 52:351–368

123

2 Tectonic setting

The Damavand Test Site is part of the central Alborz, northern Iran (Fig. 1). Several active

faults affect the central Alborz region (Berberian 1983; Berberian and Yeats 1999; Allen

et al. 2003). The main active faults in this area are the North Tehran and Mosha faults. The

Mosha fault, as one of the major active faults in the central Alborz, is characterized by its

strong historical seismicity (Ambraseys and Melville 1982) and its clear morphological

signature. Passing by the vicinity of Tehran city, this 150-km-long N100� E trending fault

represents an important potential seismic source.

3 Objectives

After a thorough study of major institutions involved in the earthquake prediction/pre-

cursor research and preparation of a major review and evaluation, NCEP defined a 5-year

plan. One major step was the definition and establishment of an earthquake precursory test

site. Several locations in Iran have been evaluated. The Damavand test site was selected

based on (i) proximity to the NCEP, (ii) high population density in the mega-city of

Tehran, (iii) low frequency but high magnitude earthquake occurrence in comparison with

other evaluated regions, (iv) proximity to major active faults with historical/instrumental

earthquake events (Fig. 1). Items (i), (ii), and (iv) had been considered as the most

important factor in the above mentioned selection criteria. The next step was the selection

of the methodology to be implemented at the test site, based on the earthquake precursors

evaluated by the IASPEI Sub-Commission on Earthquake Prediction (Wyss 1991). In

Fig. 1 Major active faults within the Damavand test site and surrounding region (modified after Hessamiet al. 2003), the test site is shown within the shaded lower center square. It also shows the location ofsprings (small circles) in the central Alborz and historical earthquake (circles with numbers that indicate theyear) occurrences in the same area (see text for more explanations); the triangle shows the location of PolrudDam

Nat Hazards (2010) 52:351–368 353

123

addition to the methodologies thus defined others were selected that are testable by other

methods.

4 Methodologies

Based on the above, the following methodologies will be used at the test site.

4.1 Seismic

The epicenters of the regional seismicity are shown in Fig. 2. The seismicity study has

been conducted previously by many such as Ambraseys (1963), Ambraseys and Melville

(1982), Berberian (1994), Berberian and Yeats (1999), and Ashtari et al. (2005).

Fig. 2 Temporary dense seismic network (dark triangle) around Mosha Fault and the other most importantfaults in Damavand test site (black continuous lines). The seismicity shown is from 1995 to 2005. Theevents concentration at the southeast of Tehran city is due to mining activities, not earthquake (Tatar 2008,personal communication

354 Nat Hazards (2010) 52:351–368

123

However, before the permanent installation of a monitoring network, a feasibility study

for optimized location identification has been conducted in the area covering the test site.

Figure 2 shows the location of the temporary seismic stations (triangles). The network

consisted of 49 short-period three-component seismometers, with sampling rate of

100 samples/s and was operated for approximately 4 months in 2006. The analysis of these

data is ongoing (Tatar 2008, personal communication). The major concentration of events

close to southeast of Tehran city (Fig. 2) is related to mining activities in this area. These

events will be subtracted from the real earthquake events based on the explosive firing

times provided by the mining companies.

It is important to mention that there are three broadband 3c seismic stations (using

VSAT communication), one short-period seismic array (using telemetric communication),

and 13 short-period seismic stations (Yamini-Fard et al. 2008; manually downloaded

regularly) within and in close proximity of the Damavand test site. It is planned to integrate

these stations also in the future data analysis for earthquake precursor proposes.

The planned permanent monitoring stations will be consisted of 25 short-period three-

component seismometers with sampling rate of 100 samples/s. It is further anticipated that

the data will be made available online for fast analysis using the VSAT communication

system of the existing National Broad-band Seismic Network operated by IIEES. In this

network, SEISAN (Havskov and Ottemoller 2005) is being used for data analysis. At the

first instance, the same analysis software will be used and later if required new software

will be developed for fast data analysis.

In addition, another methodology based on Markov process is now under development

and evaluation in cooperation with the Physics Department at Sharif University, for ana-

lyzing precursory seismic data. The method treats seismic data as a Markov process and

distinguishes real fluctuations due to the generation of anomalous seismic noise of crack

development from the background noise (Rahimi Tabar et al. 2007). It has been observed

that the Markov time scale (tcM) increases sharply short time before an earthquake, on the

order of several hours, hence providing an alarm for an impending earthquake (Rahimi

Tabar et al. 2007; Mokhtari et al. 2007b). To distinguish a false alarm from a true event, a

second quantity (Tc1) is being computed, based on the concept of extended self-similarity

of the data (Rahimi Tabar et al. 2007; Mokhtari et al. 2007b). The method with the data for

the stations in one region (i.e., estimate tcM and Tc1 for distances d \ dc) has been cali-

brated (Mokhtari et al. 2007b). So, if in a given region there is a single station, then once

the online-computed tM and T1 exceed their critical values, the alarm is turned on. If there

are several stations in the region, then once they declare that their tM and T1 have exceeded

their thresholds, the alarm is turned on. So, if after about 2 h, no earthquake has occurred,

then the expected magnitude of the incoming earthquake will be greater than Mc = 4.5 at a

distance of d \ dc (Rahimi Tabar et al. 2007). In this analysis, the vertical ground velocity

data was analyzed because, with the method described above, they provide relatively long

(on the order of several hours) and, hence, useful alarms for the impending earthquakes

(Rahimi Tabar et al. 2007).

Figure 3 illustrates ground velocity Vz(t), T1(t), and tM(t) for an earthquake, recorded at

Ashtian seismic station (this station is part of IIEES permanent online seismic network

already being connected to the above mentioned software) at a distance of about 150 km

from the epicenter. The thresholds used for this station based on the above mentioned

calibration result are, tMc = 12 and T1c = 8. Therefore, T1 and tM provided 5 h alarm for

the earthquake. It is important to note that there was no foreshock for several days before

the main event.

Nat Hazards (2010) 52:351–368 355

123

It must be pointed out that the most accurate alarms are obtained from stations that are

perpendicular to the active faults that cause the earthquake (Rahimi Tabar et al. 2007).

As mentioned above, the method is under development and efforts to improve the

system for prevention of false alarm, and magnitude determination, are ongoing. The

monitoring network mentioned here will be connected to the above mentioned software

installed at the IIEESS-NCEP central processing center for possible precursory event

detection at the Damavand test site.

Fig. 3 Vertical ground velocity Vz(t), T1(t), and tM(t) for M = 5.7 earthquake that occurred on March 13,2005, at 03:31:21 a.m. in Saravan at (27.37 N, 62.11 E) in southern Iran, recorded at Ashtian station at adistance of about 150 km from the epicenter, where tM is in number of data points (the frequency at thestation is 40 Hz), T1 is dimensionless, while Vz(t) is in ‘‘counts’’ which, when multiplied by a factor1.1382 9 10-3, is converted to lm/s. a The earthquake event. b The TM and alert. c T1 and the alert (afterRahimi Tabar et al. 2007; Mokhtari et al. 2007b)

356 Nat Hazards (2010) 52:351–368

123

It has long been suspected that changes in shear wave splitting would monitor temporal

changes of micro-crack geometry caused by changes of stress before earthquakes

(Crampin 1978; Crampin et al. 1984). This methodology has been used elsewhere, for

example, in Iceland (Crampin et al. 2003; Crampin and Gao 2006). Therefore, by gradual

increase in know-how within the NCEP, this methodology will also be examined at the

test site.

4.2 GPS stations

Recent developments in the positioning techniques, especially satellite positioning and the

appearance of Global Positioning System (GPS), have brought significant improvements to

crustal deformation studies. GPS provides valuable information in crustal behavior and the

deformation rate due to stresses. Around the test site, a GPS network with 68 elements has

been designed. Figure 4 shows the location of the planned GPS stations and existing

permanent GPS stations (white circles and filled triangles, respectively). Most stations will

be seated on hard rock, though some can only be located on soft sediments due to geodetic

requirements. The objectives of this GPS network are (i) geodynamic studies, (ii) strain

analysis and observing geodetic precursors, (iii) inter-seismic studies and viscoelastic

modeling, and (iv) co-seismic studies and fault modeling.

The GPS network will allow for (i) objective function selection, (ii) geologic studies of

the test site, (iii) GPS network design based on selected objective function and geological

constraints, (iv) network implementation and observation, (v) data processing, (vi) mod-

eling, (vii) integration with seismic and geologic data for precursor identification.

Fig. 4 Location of the designed GPS network based on geological constraints and the major faults (whitecircles are planned GPS locations and the black triangles are existing permanent GPS stations)

Nat Hazards (2010) 52:351–368 357

123

4.3 Magnetic

Magnetic data have played a significant role in earthquake prediction studies over the

decades. Tectonomagnetic monitoring using different network layout with varying number

of magnetometers has been successfully conducted in many seismically active provinces

(e.g., Rikitake et al. 1980; Abdullabekov 1987; Johnston 1989; Zlotnicki and Le Mouel

1990; Kuznetsova and Maksimchuk 1994; Shapiro et al. 1994; Dyadkov et al. 1999;

Meloni et al. 2004; Chen et al. 2004; Di Mauro et al. 2007).

This goal is not easy to achieve due to the natural diurnal variations of the local

magnetic field, induced by the ionospheric current vortices and magnetic storms, and due

to the ‘‘cultural’’ noise arising from human activities. Detection of low tectonomagnetic

anomalies therefore requires high-precision, stable instrumentation, and special processing

(filtering) techniques to separate the signals from noise (Davis et al. 1981; Dyadkov 1985).

However, it is important to note that the cultural noise at the Damavand test site is

relatively well characterized. Diurnal variations are routinely obtained from continuous

magnetic observatory records, and the intensity distributions of these daily variations

across the hemisphere are known. Hence, it is believed that this type of study can be

successfully conducted at the Damavand test site. The objectives of acquiring periodical/

continuous magnetic data at the Damavand test site includes observing magnetic precur-

sors and carrying out inter-seismic studies along with co-seismic studies and fault

modeling, and finally integration of the result with other methodology for precursory

studies.

The study procedures are (i) objective function selection, (ii) magnetic network design

based on selected objective function and geological constraints, (iii) network implemen-

tation, (iv) observations, (v) data processing and modeling, and (vi) integration with

seismic, GPS, and geologic data.

4.3.1 Magnetic data gathering

As mentioned above, other methods have gained some attention as possible pre-earthquake

indicator, in particular changes in the permanent local magnetic field (for example, Chen

et al. 2004; Liu et al. 2006). A feasibility study to establish a magnetic baseline for the area

has been conducted across the Mosha fault (Mokhtari et al. 2008a). Figure 5 shows the

location of the magnetic profiles so far obtained. Figure 6 shows a magnetic anomaly along

PC13; the values indicate that large anomalies occur, which are large enough to be detected

by magnetometers to be installed at the test site. In addition, the large anomaly across the

Mosha fault confirms its location, which had already been defined by direct observations.

4.4 Morphostructural zoning (MZ) method

Based on a cooperative agreement between IIEPT&MG RAS (Russian Academy of Sci-

ence) and NCEP of IIEES, the morphostructural zoning (MZ) method has been applied to

the Alborz Mountains aimed at (i) the compilation of the morphostructural map for the

Alborz Mountains at the scale of 1:500,000, (ii) the correlation of the delineated nodes with

the M C 6.0 earthquakes, and (iii) the parameterization of the delineated nodes (Gorshkov

et al. 2007; Mokhtari et al. 2007a).

The main objective is to identify the nodes in the Alborz Mountains where the epi-

centers of strong earthquakes may be situated. For this purpose, the methodology

developed by IIEPT&MG that has been applied to many seismic regions of the world for

358 Nat Hazards (2010) 52:351–368

123

the identification of seismogenic nodes has been employed (Bhatia et al. 1992; Gelfand

et al. 1972, 1976; Gorshkov et al. 1991, 2000, 2002, 2003, 2004; Gvishiani et al. 1986,

1987, 1988). Recent earthquakes in each of the regions studied have proven the reliability

of the results obtained. As Gorshkov et al. 2003 demonstrated, 90% of the post-earthquake

events with relevant magnitudes occurred at the nodes and 84% of the post-earthquakes

events occurred at the nodes recognized as prone to strong earthquakes.

Two principal steps compose the methodology. The first step is the delineation of the

morphostructural nodes by the MZ method (Alexeevskaya et al. 1977; Rantsman 1979;

Gorshkov et al. 2003). The second is the classification of all mapped nodes, by the pattern

recognition algorithm CORA-3 (Gelfand et al. 1976; Gvishiani et al. 1988; Gorshkov et al.

2003) into nodes where earthquakes with magnitude exceeding a certain threshold are

possible and nodes where only earthquakes with smaller magnitude may happen.

Fig. 5 Location of geomagnetic profiles acquired within the test site

Fig. 6 Magnetic anomaly profilealong PC13 (see for locationFig. 5) after all requiredcorrection. The profile crosses theMosha fault where majoranomaly due to the fault could beobserved

Nat Hazards (2010) 52:351–368 359

123

At the first stage, a map of the Alborz Mountains has been compiled at the scale of

1:500,000 using the MZ method (Alexeevskaya et al. 1977; Rantsman 1979; Gorshkov

et al. 2003). The map shows the hierarchical block structure of these regions, the network

of boundary zones separating blocks, and the loci of the nodes. Nodes are the specific

structures formed at the intersections of boundary zones.

According to the principles of MZ, the Alborz is considered to be a single mountain

zone created by the Alpine orogenesis (Alavi 1996; Khain 2000). First rank boundaries

shown in Fig. 7 separate the Alborz from the surrounding large-scale geotectonic domains.

The Alborz region is decomposed into 17 mega-blocks recognized by differences in ele-

vation and orientation of individual blocks composing the mountain belt. Second rank

lineaments dissect the delineated mega-blocks. The longitudinal segmentation of the Al-

borz into mega-blocks is controlled by longitudinal second rank lineaments that correspond

to the prominent faults shown on geological map of Iran.

The present-day Alborz is very heavily dissected. Across the mountain chain, the

quantitative index of the topography changes sharply over short distances. Therefore, MZ

allows for a dense network of third rank lineaments. Specifically, this is true for the eastern

part of the Alborz. These lineaments have been traced along steep scarps on the slopes of

the ridges and the rectilinear fragments of river valleys that usually are fault-dominated in

young mountains. In total, with the MZ method, 134 intersections of lineaments in the

Alborz region have been outlined (Fig. 7). Each of them has been considered as a node.

Figure 8 presents selected earthquakes and shows the spatial correlation between the

events with M C 6.0 and morphostructural nodes delineated in the Alborz region.

Based on the results of this study, it has been confirmed that there is a high seismic

potential of the Alborz region and the importance of topography in the identification of

earthquake-prone areas. This study is ongoing and in the final result, geological/tectonic

field work, gravity, and other geophysical data will be included.

Fig. 7 Morphostructural map of the Alborz region (all ranks). Thick lines indicate the lineaments of the firstrank; medium lines show lineaments of the second rank; thin lines depict the lineaments of the third rank.Continues lines depict the longitudinal lineaments, while the discontinuous lines represent the transverselineaments (Mokhtari et al. 2007a)

360 Nat Hazards (2010) 52:351–368

123

4.5 Water level changes

Changing water levels in deep wells are recognized by IASPEI as a significant precursor to

earthquakes (Wyss 1991). There are many examples in the literature where water level

changes have been associated with earthquakes, either as precursors or as co-seismic and

post-seismic events (Roeloffs 1988; Contadakis and Asteriadis 2001; Koizumi et al. 2005;

Kumpel 1992; Wang 1984; Crampin et al. 2003). For example, about 3 months before the

Longling earthquakes, the water level at a well at Xiaguan, Yunnan Province, began to

lower by centimeters (Rikitake 1982). Some 10 days after the water level increased, the

earthquake occurred (Rikitake 1982).

It has been demonstrated that seismic waves can induce large water level fluctuations in

wells. Large amplitude surface seismic waves, such as Raleigh Waves, force the particles

of the rock near the surface to move in an elliptical orbit and thus the aquifer layer also is

affected, which in turn results in the water level fluctuation in the well. Thus, if appropriate

measuring instruments are used, the water level in wells can be used for recording distant

earthquakes. In essence, a well can act as a seismograph by recording the passage of the

surface waves through the aquifer and amplifying the amplitude of these waves, much like

a seismograph does. Thus, many major earthquakes throughout the world have produced

water level changes.

Not only do water level changes occur following an earthquake, they also precede most

earthquakes. Water wells are very sensitive to various earth processes such as earth tides,

tilting of the crust, the dilatancy of cracks, pore space in the rock, and aseismic creep,

particularly if these wells are in the close proximity to an active fault. By drilling water

wells at carefully selected sites and by measuring water level and water quality, the

information can be used as earthquake precursor, particularly if it is used in conjunction

with a dense network of other instruments such as tilt meters and creep meters. Thus, actual

Fig. 8 Recognized seismogenic nodes prone to earthquakes with M C 6.0. The large dots define theearthquake events and the circles are the nodes defied based on morphostructural zonation (Mokhtari et al.2007a)

Nat Hazards (2010) 52:351–368 361

123

pre-seismic processes and precursory fluctuations in water levels can give a clear indica-

tion of strain build-up along a particular seismic fault.

A number of deep wells in China are reported to be extremely sensitive to changes in

pressure and can reliably detect earthquakes occurring halfway round the world. This

observed sensitivity is probably due to their being quite protected from surface noise

(rainfall, seasonal effects, and others). As a result, China relies a great deal on deep wells

for earthquake prediction. Indeed, over 100 research wells in excess of 1,000 m deep have

been drilled solely for earthquake prediction purposes. In these wells, water levels are

continually monitored to ±0.5 cm and temperatures to ±0.01�C. Japan also relies to some

extent on such wells—some 93 wells are monitored for earthquakes. In addition, changes

in water level and water pressure variations in well on Island of Flatey, Iceland, have been

reported before an impending earthquake by Crampin et al. (2003).

In the monitoring of water levels in deep wells, care must be taken to correct for water

extraction from the aquifer. In many parts of the planet, the water table is falling due to

water extraction for drinking and irrigation. It is quite possible that such drops could be

mistaken for a long-term seismic precursor.

4.5.1 An observation of the Bam earthquake in Iran

The Bam earthquake occurred in the Kerman Province of Iran on 26 December 2003,

with magnitude 6.7. It occurred in an area considered to be relatively ‘‘non-seismic’’ and

resulted in the huge loss of life and property. Based on data observed at the Polrud Dam

(Fig. 1) located in the north of Iran, a gradual change in the water level before three

earthquakes, including the Bam event, has been identified. The piezometer reading has

been done once a week. The precipitation rate from two weather stations nearby con-

firmed that the water level variations in the observation location are not the result of the

changes due to precipitation. Figure 9 shows the earthquake correspondence with the

gradual water level changes and the occurrence of the Bam earthquake at a distance of

1,258 km. So far, based on the information available, we have no explanation for this

observation and we have no indication that can support this phenomena as precursory

event. However, post-earthquake observation has shown this correspondence and work is

continuing in this regard. In addition, a few similar events have been observed in other

Fig. 9 Water level changes in association with the Bam and two other pre-Bam earthquake events (close tothe Bam earthquake) based on piezometer observation at the Polrud Dam observation site. Readings of thepiezometer were taken once a week (see for location Fig. 1)

362 Nat Hazards (2010) 52:351–368

123

locations in Iran and other parts of the world that provide some encouragement for

applying this method.

Based on the above, it is planned to monitor a few existing deep exploration and

observation wells at the Damavand test site for the study of oscillations in water levels,

micro-temperature signals, ground water chemistry, electrical conductivity, TDS, pH,

turbidity, and dissolved gases such as radon, CO2, He, CH4, and H2 as earthquake pre-

cursors. Study of physical and chemical properties of ground water especially changes in

water levels in observation wells can be useful for recognition of under inductive stress

regions (compression or tension) due to the occurrence of near earthquakes and also long-

term seismotectonic anomalies in seismic regions. It is further anticipated that within the

test site, additional deep water wells will be drilled and monitored.

4.5.2 Springs

The study of physical and chemical properties of springs (changes of outflow, temperature,

chemistry, dissolved gas) as earthquake precursors represents yet another methodology that

will be used at the Damavand test site. To achieve this objective, the springs with high

water fluctuations, deep sources, and high debit have been selected. Several major springs

around Tehran will be included in this study.

Figure 1 show the main springs in the Alborz region alongside the locations of large

earthquakes and locations of springs. It is an interesting observation that, with exception of

the Roudbar region, there is a correlation between the occurrence of large earthquakes and

the area where no spring has been identified. It has to be emphasized that this is an

observation and requires further investigations.

4.6 Thermal anomalies

Since the late 1980s and early 1990s, non-stationary areas of enhanced infrared (IR)

emission, linked to impending earthquake activity, have been recognized in night-time

satellite images (Gornyi et al. 1988; Qiang et al. 1990, 1991; Srivastav et al. 1997) with

apparent land surface temperature variations on the order of 2–4�C and occasionally

higher. These areas have become known as ‘‘thermal anomalies.’’ Their cause has

remained enigmatic (Cui et al. 1999; Srivastav et al. 1997; Tronin 2000; Tronin 2002),

though recent work provides more evidence that the effect is real (Tronin et al. 2004). The

rapidity of the temperature variations rules out the flow of Joule heat from sources deep

below. Several processes have been invoked to account for the observations: (i) emanation

of warm gases (Gornyi et al. 1988), in particular CO2 emanation, causing a ‘‘local

greenhouse’’ effect (Qiang et al. 1991; Tronin 1999, 2002); (ii) rising well water levels and

changing moisture contents in the soil; (iii) near-ground air ionization due to enhanced

radon emission leading to the condensation of water vapor from the atmosphere and the

release of latent heat (Liperovsky et al. 2005). None of the explanations listed in (i)–(iii)

can adequately account for the lateral extent of the ‘‘thermal anomalies,’’ for the rapidity

with which the areas of enhanced IR emission appear and disappear. Since it is highly

unlikely that (i)–(iii) are simultaneously valid, we are faced with two alternatives: either

the ‘‘thermal anomalies’’ are an artifact after all or some as yet unrecognized physical

process is responsible for the enhanced IR emission from the ground. Recently, Freund

(2007b, c) drew attention to a non-classical emission process, demonstrated in the labo-

ratory (Freund et al. 2007), i.e., due to the flow of stress-activated electronic charge carriers

Nat Hazards (2010) 52:351–368 363

123

from deep within the crust to the Earth’s surface. At the surface, these charge carriers

recombine forming vibrationally highly excited states. Those states radiatively de-excite

emitting a series of IR bands in the spectral region of the thermal IR emission, which may

be confused with a temperature-induced IR emission.

However, based on data available, few successful events have been recognized before

earthquake occurrences in Iran, similar to other locations in other parts of the world.

Figure 10 shows alerts based on thermal anomalies (stars) and earthquake events (circles)

during 2006 with magnitude larger than 4 in Iran (Mokhtari et al. 2008b). There have

been 23 alerts (M [ 4) during 2006, from which 15 events have occurred within a 100-

km radius of the epicenter and within 2 months period of predicted time. It is important

to note the above mentioned results require further investigation and more rigorous

examinations.

Fig. 10 Alerted events based on thermal anomalies (stars) and earthquake events (circles) occurred during2006 with magnitude larger than 4 in Iran

364 Nat Hazards (2010) 52:351–368

123

5 Conclusions and future works

A detailed review of earthquake prediction/precursory activities worldwide has been

conducted by NCEP. Based on this study, a document has been prepared which contains

the plan to be implemented over the next 5 years. After review of five different sites within

northern Iran, the Damavand test site was selected based on proximity to the capital of Iran

with its high population density, low frequency but high magnitude earthquake occurrence,

and the existence of active faults with historical earthquake events. The methodologies to

be implemented are based on the IASPEI evaluation report and NCEP’s capabilities.

In this context, extensive geomagnetic measurements have already been conducted to

test the magnetic response of the area and also to better understand the Mosha fault. The

MZ methodology, although conducted over the whole Alborz region, gave confirmation of

the high seismic potential of the test site, which in turn confirmed the suitability of the test

site. The temporary seismic network analysis is ongoing for better definition of permanent

seismic monitoring station locations. Water level fluctuation in other parts of Iran confirms

its potential as earthquake precursor but requires further analysis and calibration to be fully

implemented.

It is anticipated that the establishment of the Damavand test site in a seismically active

zone such as the central Alborz region can contribute to an organized research activity

within the earthquake precursory disciplines.

Other additional important aspects that would require further attention are (i) a pre-

cursor warning system that can disseminate the precursory/prediction result to the decision-

making body or public and (ii) an early warning system for fast, intelligent automatic

recognition of a large earthquake from nearby records of the first few seconds of its P

wave-train, the system then instantaneously transmitting the alert electronically to stop or

shut down more distant critical systems such as power stations, transport, before the most

damaging surface waves arrive. Early warning systems are already being operated in some

parts of the world such as EWS in USA, IERREWS in Turkey, and UrEDAS in Japan. In

this regard, design and implementation of a single, or preferably multi-hazard, early

warning system based on multidisciplinary results is an essential part of a precursor/

prediction scheme.

It is of utmost importance for the success of the Damavand test site that it draws

international support both in terms of technologies and examination of theories. Interna-

tional cooperation is invited and encouraged.

Acknowledgments This work would not have been achieved without kind and generous support of Prof.M. Ghafoury-Ashtiany, Dr. F. Yami-fard, Dr. K. Hessami, Mrs. L. Mahshadnia, Mrs. P. Moboyen, Mr. M.Shierzaei, Mrs. M. Akbari, Mr. A.M. Asgari, Dr. M. Tatar, Dr. M. R. Rahimi Tabar, and Dr. I. Abdollahie-Fard, whom I am very thankful. I would like to express my sincere thanks to Dr. F. Freund and twoanonymous reviewers for critically reviewing the manuscript and giving valuable advice not only inimproving the text but also correcting them technically.

References

Abdullabekov KN (1987) Electromagnetic phenomena in the earth’s crust. FAN Press, pp 3–231Alavi M (1996) Tectonostratigraphic synthesis and structural style of the Alborz mountain system in

northern Iran. J Geodyn 21(1):1–33. doi:10.1016/0264-3707(95)00009-7Alexeevskaya MA, Gabrielov AM, Gvishiani AD, Gelfand IM, Rantsman EY (1977) Formal morpho-

structural zoning of mountain territories. J Geophys 43:227–233

Nat Hazards (2010) 52:351–368 365

123

Allen MB, Ghassemi MR, Shahrabi M, Qorashi M (2003) Accommodation of late Cenozoic obliqueshortening in the Alborz range, northern Iran. J Struct Geol 25:659–672. doi:10.1016/S0191-8141(02)00064-0

Ambraseys NN (1963) The Buyin-Zara (Iran) earthquake of September, 1962, a field report. Bull SeismolSoc Am 53:705–740

Ambraseys NN, Melville CP (1982) A history of Persian earthquakes. Cambridge University Press, Cam-bridge, 219 pp

Ashtari M et al (2005) Microseismicity in the region of Tehran. Tectonophysics 395:193–208. doi:10.1016/j.tecto.2004.09.011

Berberian M (1983) The southern Caspian: a compressional depression floored by a trapped, modifiedoceanic crust. Can J Earth Sci 20:163–183

Berberian M (1994) Natural hazard and the first earthquake catalogue of Iran. In: Historical hazards in Iranprior to 1900, vol 1. IIEES, Iran, 605 pp

Berberian M, Yeats RS (1999) Patterns of historical earthquake rupture in the Iranian plateau. Bull SeismolSoc Am 89:120–139

Bernard P et al (1997) From precursors to prediction: a few recent cases from Greece. Geophys J Int131:467–477. doi:10.1111/j.1365-246X.1997.tb06590.x

Bhatia SC, Chetty TRK, Filimonov M, Gorshkov A, Rantsman E, Rao MN (1992) Identification of potentialareas for the occurrence of strong earthquakes in Himalayan arc region. Proc Indiana Acad Sci (EarthPlanet Sci) 101(4):369–385

Chen CH, Liu JY, Yen HY, Zeng X, Yeh YH (2004) Geomagnetic total field and occurrences of earthquakesin Taiwan. TAO 15(3):361–370

Contadakis ME, Asteriadis G (2001) Hydrologic changes as possible earthquake precursor in Greece. NatHazards 23(1):29–47. doi:10.1023/A:1008176804533

Crampin S (1978) Seismic wave propagation through a cracked solid: polarization as a possible dilatancydiagnostic. Geophys J R Astron Soc 53:467–496

Crampin S, Gao Y (2006) A review of techniques for measuring shear-wave splitting above small earth-quakes. Phys Earth Planet Int 159:1–14. doi:10.1016/j.pepi.2006.06.002

Crampin S, Evans R, Atkinson BK (1984) Earthquake prediction: a new physical basis. Geophys J R AstronSoc 76:147–156

Crampin S, Chastin S, Gao Y (2003) Shear-wave splitting in a critical crust: III. Preliminary report of multi-variable measurements in active tectonics. J Appl Geophys 54:265–277. doi:10.1016/j.jappgeo.2003.01.001

Cui C et al (1999) Monitoring the thermal IR anomaly of Zhangbei earthquake precursor by satellite remotesensing technique. Paper presented at ACRS

Davis PM, Jackson DD, Searls CA, McPherron RL (1981) Detection of tectonomagnetic events usingmultichannel predictive filtering. J Geophys Res 86(B3):1731–1737. doi:10.1029/JB086iB03p01731

Di Mauro D, Lepidi S, Di Persio M, Meloni A, Palangio P (2007) Update on monitoring of magnetic andelectromagnetic tectonic signals in Central Italy. Ann Geophys 50(1):51–60. http://hdl.handle.net/2122/3374

Dyadkov PG (1985) Linearity of secondary effects of geomagnetic variations in anomalous fields in the caseof modulus measurements. Sov Geol Geophys 26(4):129–136

Dyadkov PG, Mandelbaum MM, Tatkov GI, Larionov VA, Zhirova NV, Mikheev OA, Nizamutdinov RS,Chebakov GA (1999) Development of seismotectonic and pre-seismic processes in the central Baikalrift zone, from tectonomagnetic observations. Russ Geol Geophys 40(3):343–355

Freund FT (2007a) Understanding pre-earthquake signals. In: 5th International conference on seismologyand earthquake engineering, section EP, Tehran, 13–16 May 2007

Freund FT (2007b) Pre-earthquake signals—part I: deviatoric stresses turn rocks into a source of electriccurrents. Nat Hazards Earth Syst Sci 7:535–541

Freund FT (2007c) Pre-earthquake signals—part II: flow of battery currents in the crust. Nat Hazards EarthSyst Sci 7:1–6

Freund FT, Takeuchi A, Lau BWS, Al-Manaseer A, Fu CC, Bryant NA, Ouzounov D (2007) Stimulatedthermal IR emission from rocks: assessing a stress indicator. eEarth 2:1–10

Gelfand I, Guberman S, Keilis-Borok V, Rantsman E, Izvekova M (1972) Criteria of high seismicity deter-mined by pattern recognition. Tectonophysics 13(1/4):415–422. doi:10.1016/0040-1951(72)90031-5

Gelfand I, Guberman S, Keilis-Borok V, Knopoff L, Press F, Rantsman E, Rotwain I, Sadovsky A (1976)Pattern recognition applied to earthquake epicentres in California. Phys Earth Planet Inter 11:227–283.doi:10.1016/0031-9201(76)90067-4

Gokhberg MB, Morgounov VA, Pokhotelov OA (1995) Earthquake prediction seismo-electromagneticphenomena. Gordon and Breach Publishers, London, 193 pp

366 Nat Hazards (2010) 52:351–368

123

Gornyi VI, Salman AG, Tronin AA, Shilin BB (1988) The Earth’s outgoing IR radiation as an indicator ofseismic activity. Proc Acad Sci USSR 301:67–69

Gorshkov A, Zhidkov M, Rantsman E, Tumarkin A (1991) Morphostructures of the Lesser Caucaus andsites of earthquakes, M C 5.5. Izv USSR Acad Sci Phys Earth 6:30–38 (in Russian)

Gorshkov A, Kuznetsov I, Panza G, Soloviev A (2000) Identification of future earthquake sources in theCarpatho-Balkan orogenic belt using morphostructural criteria. Pure Appl Geophys 157:79–95. doi:10.1007/PL00001101

Gorshkov A, Panza G, Soloviev A, Aoudia A (2002) Morphostructural zonation and preliminary recognitionof seismogenic nodes around the Adria margin in peninsular Italy and Sicily. JSEE 4(1):1–24

Gorshkov A, Kossobokov V, Soloviev A (2003) Recognition of earthquake prone areas. In: Kelis-Borok V,Soloviev A (eds) Nonlinear dynamics of the lithosphere and earthquake prediction. Springer, Hei-delberg, pp 235–320

Gorshkov A, Panza G, Soloviev A, Aoudia A (2004) Identification of seismogenic nodes in the Alps andDinarides. Boll Soc Geol Ital 123:3–18

Gorshkov A, Mokhtari M, Piotrovskaya E, Shahpasanzadeh M (2007) Identification of earthquake proneareas in the Alborz Region. In: 5th International conference on seismology and earthquake engi-neering, section EP, Tehran, 13–16 May 2007

Gvishiani A, Gorshkov A, Kossobokov V, Rantsman E (1986) Morphostructures and earthquake-prone areasin the Greater Caucasus. Izv USSR Acad Sci Phys Earth 9:15–23 (in Russian)

Gvishiani A, Gorshkov A, Kossobokov V, Cisternas A, Philip H, Weber C (1987) Identification of seis-mically dangerous zones in the Pyrenees. Ann Geophys 5B(6):681–690

Gvishiani A, Gorshkov A, Cisternas A, Rantsman E, Soloviev A (1988) Identification of earthquake-prone-areas in the regions of moderate seismicity. Nauka, Moscow, 189 pp (in Russian)

Havskov J, Ottemoller L (2005) SEISAN: the earthquake analysis software. University of Bergen, NorwayHessami K, Jamali F, Tabassi H (2003) Major active faults of Iran. IIEES, TehranHolliday JR et al (2005) Earthquake forecasting and its verification. Nonlinear Process Geophys 12:965–977Hough SE (2002) Earthshaking science: what we know (and don’t know) about earthquakes. Princeton

University Press, Princeton, 272 ppJohnston MJS (1989) Review of magnetic and electric field effects near active faults and volcanoes in USA.

Phys Earth Planet Inter 57:47–63. doi:10.1016/0031-9201(89)90213-6Kagan YY (1997) Are earthquakes predictable? Geophys J Int 131:505–525Kanamori H (1996) A seismologist’s view of VAN. In: Sir Lighthill J (ed) Earthquake prediction from

seismic electric signals. World Scientific, Singapore, pp 339–345Keilis-Borok V (2002) Earthquake prediction: state-of-the-art and emerging possibilities. Annu Rev Earth

Planet Sci 30:1–33Keilis-Borok V, Soloviev AA (eds) (2003) Nonlinear dynamics of the lithosphere and earthquake prediction.

Springer, Heidelberg, 337 ppKhain VE (2000) Tectonics of continents and oceans. Scientific World, Moscow, 500 pp (in Russian)Knopoff L (1996) Earthquake prediction: the scientific challenge. Proc Natl Acad Sci USA 93:3719–3720.

doi:10.1073/pnas.93.9.3719Koizumi N, Takahashi M, Matsumoto N, Sato T, Ohtani R, Kitagawa Y (2005) Hydrological research for

earthquake prediction: trial for detection of preseismic crustal deformation from groundwater changes.J Seismol Soc Jpn 58(3):247–258

Kumpel H (1992) About potential of wells to reflect stress variations within inhomogeneous crust. Tec-tonophysics 211:317–336. doi:10.1016/0040-1951(92)90068-H

Kuznetsova VG, Maksimchuk VE (1994) Tectonomagnetic investigations for study of structure and recentgeodynamics of lithosphere of the Carpatian region. J Geophys Geodyn 1(1):31–42

Liperovsky VA et al (2005) On the possible influence of radon and aerosol injection into the atmosphere andionosphere before earthquakes. EGU Spring Meeting, Vienna

Liu JY, Chen CH, Chen YI, Yen HY, Hattori, Yumoto K (2006) Seismo-geomagnetic anomalies andM [ 5.0 earthquakes observed in Taiwan during 1988–2001. Phys Chem Earth 31:215–222

Lomnitz C (1994) Fundamentals of earthquake prediction. Wiley, New York, 326 ppMasood E (1995) Greek earthquake stirs controversy over claims for prediction method. Nature 375:617 ppMeloni A, Mauro DD, Lepidi S, Mele G, Palangio P (2004) Tectonomagnetic and VLF electromagnetic

signals in central Italy. Ann Geophys 47(1):29–37Milne J (1899) Earthquakes and other earth movements, 4th edn. D Appleton Co, New York, 376 ppMokhtari M, Groshkov A, Shahpasandzadeh M (2007a) Identification of seismic pattern in the central

Alborz using morphostructural zone method for earthquake of larger or equal 6. Internal technicalpublication of IIEES

Nat Hazards (2010) 52:351–368 367

123

Mokhtari M et al (2007b) Calibration of earthquake early warning system using broad band seismic networkof IIEES. Internal technical publication of IIEES

Mokhtari M et al (2008a) Magnetic survey of the Mosha fault region—central Alborz. Internal technicalpublication of IIEES

Mokhtari M et al (2008b) Introduction to earthquake prediction. Naghoos Andisheh, Tehran, 152 pp (in Farsi)Park SK (1997) Electromagnetic precursors to earthquakes: a search for predictors. Sci Prog 80:65–82Qiang Z-J, Xu X-D, Dian C-D (1990) Abnormal infrared thermal of satellite-forewarning of earthquakes.

Chin Sci Bull 35:1324–1327Qiang Z-J, Xu X-D, Dian C-D (1991) Thermal infrared anomaly—precursor of impending earthquakes.

Chin Sci Bull 36:319–323Rahimi Tabar MR, Sahimi M, Kaviani K, Allamehzadeh M, Peinke J, Mokhtari M Vesaghi M et al (2007)

Short-term prediction of medium and large-size earthquakes based on Markov and extended self-similarity analysis of seismic data. In: Modeling critical and catastrophic phenomena in geoscience: astatistical physics approach. Lecture notes in physics, vol 705, Springer Verlag, Berlin, pp 281–301

Rantsman EY (1979) Morphostructure of mountain regions and sites of earthquakes. Nauka, Moscow, 179pp (in Russian)

Rikitake T (1982) Earthquake forecasting and warning. Center for Academic Publications, TokyoRikitake T, Honkura Y, Tanaka H, Ohshiman N, Sasay Y, Ishikawa Y, Koyama S, Kawamura M, Ohchi K

(1980) Changes in the geomagnetic field associated with earthquakes in the Izu Peninsula. Jpn JGeomagn Geoelectr 32(12):721–739

Roeloffs EA (1988) Hydrologic precursors to earthquakes: a review. Pure Appl Geophys PAGEOPH 126(2–4):177–209. doi:10.1007/BF00878996

Rundle JB et al (2003) Statistical physics approach to understanding the multiscale dynamics of earthquakefault systems. Rev Geophys 41:1019–1049. doi:10.1029/2003RG000135

Shapiro V, Muminov M, Khadzhyev T, Abdullabekov K (1994) Magnetic field variations of crustal originmeasured in the Fergana Valley of Usbekistan, reflecting seismotectonic dynamics. In: Hayakawa M,Fujinawa Y (eds) Electromagnetic phenomena related to earthquake prediction. Terra ScientificPublishing Co, Tokyo, pp 43–49

Shebalin P et al (2004) Advance short-term prediction of the large Tokachi-oki earthquake, September 25,2003, M = 8.1: a case history. Earth Planets Space 56:715–724

Srivastav SK, Dangwal M, Bhattachary A, Reddy PR (1997) Satellite data reveals pre-earthquake thermalanomalies in Killari area, Maharashtra. Curr Sci 72:880–884

Sykes L, Shaw B, Scholz C (1999) Rethinking earthquake prediction. Pure Appl Geophys 155:207–232. doi:10.1007/s000240050263

Tributsch H (1984) When the snakes awake: animals and earthquake prediction. MIT Press, Cambridge, 264 ppTronin AA (ed) (1999) Satellite thermal survey application for earthquake prediction. Terra Scientific

Publishing Co, Tokyo, pp 717–746Tronin AA (2000) Thermal satellite data for earthquake research. Paper presented at IGARSS 2000. IEEE

2000 international geoscience and remote sensing symposium. Taking the pulse of the planet: the roleof remote sensing in managing the environment, IEEE, Honolulu

Tronin AA (2002) Atmosphere-lithosphere coupling: thermal anomalies on the earth surface in seismicprocess. In: Hayakawa M, Molchanov OA (eds) Seismo-electromagnetics: lithosphere-atmosphere-ionosphere coupling. Terra Scientific Publishing Co, Tokyo, pp 173–176

Tronin AA, Molchanov OA, Biagi PF (2004) Thermal anomalies and well observations in Kamchatka. IntJ Rem Sens 25:2649–2655

Turcotte DL (1991) Earthquake prediction. Annu Rev Earth Planet Sci 19:263–281. doi:10.1146/annurev.ea.19.050191.001403

Uyeda S (1998) VAN method of short-term earthquake prediction shows promise. EOS Trans Am GeophysUnion 79:573–580. doi:10.1029/98EO00417

Wang C (1984) Ground-water studies for earthquake prediction in China. Pure Appl Geophys PAGEOPH122(2–4):215–217

Wyss M (1991) Evaluation of proposed earthquake precursors. American Geophysical Union Mongraph,Washington, DC, 94 pp

Wyss M, Dmowska R (eds) (1997) Earthquake prediction—state of the art. Birkhauser Verlag, Basel, 272 ppYamini-Fard F, Moradi AS, Hosseini M, Norouzi R (2008) Seismicity of Tehran city region and its close

vicinity based on Tehran city seismic network (TCSN) data. Geosciences, Tehran (submitted forpublication)

Zlotnicki J, Le Mouel JL (1990) Possible electrokinetic origin of large magnetic variations at La Fournaisevolcano. Nature 343(6259):633–636. doi:10.1038/343633a0

368 Nat Hazards (2010) 52:351–368

123