**6. Discussion of the results - assessment of energy (maximum magnitude) of paleoearthquakes by the length of the seismogenic rupture**

We have collected all absolute dates of paleoearthquakes occurred along the Talas-Fergana Fault obtained by us and combined them with the data obtained earlier by V.S. Burtman et al. (1987, 1996) and V.G. Trifonov et al. (1990, 1992) in table 4 and in Fig. 41. Totally there were collected 55 dates in 14 localities appurtenant to Late Pleistocene – Holocene.

An analysis of Fig. 41 has led us to a conclusion that the Talas-Fergana Fault zone can be generally divided onto three chains according to peculiarities of the lateral distribution of earthquakes. The first – north-western chain starts from the Kurkureusu river valley in the far west of Kyrgyzstan and stretches to about Sary-Bulak river valley in the Ketmen-Tyube depression. The second – central chain stretches toward south-east from the Sary-Bulak river valley to the Urumbash River (Kazarman depression) inclusively. Third – south-western

It is clear that such values of displacements cannot be attributed to only one earthquake. Indeed absolute dates (table 1) testifies on minimum eight earthquakes occurred in 15800,

In tie with mainly strike-slip movements along the fault, often in its zone there are formed specific deformation forms: so-called tectonic swells (Fig. 40), which are like a barrier ridges for consequent drainage and near-fault depressions – places of accumulations of fine material from adjacent slopes. In those depressions one has to conduct pit excavations for

Fig. 40. A view toward north-west along the Talas-Fergana Fault. A region of the Semiz river mouth. Here the springs – tributaries o of the river flowing down toward south-west are dammed by the tectonic swells (shown by pointers) serving as barrier or "shatter" ridges

**6. Discussion of the results - assessment of energy (maximum magnitude) of** 

We have collected all absolute dates of paleoearthquakes occurred along the Talas-Fergana Fault obtained by us and combined them with the data obtained earlier by V.S. Burtman et al. (1987, 1996) and V.G. Trifonov et al. (1990, 1992) in table 4 and in Fig. 41. Totally there

An analysis of Fig. 41 has led us to a conclusion that the Talas-Fergana Fault zone can be generally divided onto three chains according to peculiarities of the lateral distribution of earthquakes. The first – north-western chain starts from the Kurkureusu river valley in the far west of Kyrgyzstan and stretches to about Sary-Bulak river valley in the Ketmen-Tyube depression. The second – central chain stretches toward south-east from the Sary-Bulak river valley to the Urumbash River (Kazarman depression) inclusively. Third – south-western

were collected 55 dates in 14 localities appurtenant to Late Pleistocene – Holocene.

**paleoearthquakes by the length of the seismogenic rupture** 

4590, about 3955, about 3095, about 2635, 2230,1940, 1720 years ago.

samples collection for absolute dating.

for consequent drainage.



Table 4. Radiocarbon dates of samples collected from displaced gullies along the Talas-Fergana Fault (by data of Korjenkov et al., 2009, 2010, as well as V.S. Burtman et al., 1987, 1996 and Trifonov et al., 1990, 1992)

chain starts from Kyldau river valley and ends in a region of southern state border of Kyrgyz Republic. We understand whole conventionality of our division, especially without materials on the Talas-Fergana Fault from Kazakhstan and China.

Because of uneven investigations and bareness of existed materials one has to talk very carefully on segmentation of the Talas-Fergana Fault zone. According to existing data we select 13 segments: 3 – in the northwestern chain of the fault, 5 or 6 in the central chain and 4 or 5 in southeastern chain (See Fig. 41).

An analysis and comparison of materials of Table 4 and Fig. 41 have allowed us to reveal 18 paleoseismic events, 17 of which occurred in the second half of Holocene (Table 5 and Fig. 42). We assessed also distances between localities, where there were determined absolute ages of the seismogenic displacements, occurred (supposedly) during one seismic event. We conditionally accepted these distances as minimum lengths of the seismogenic ruptures.

Some of the extreme values of rupture lengths, such as 270 km and 220 km for earthquakes occurred in 4530 and 1980 years ago, provoke a natural doubt. Although such length of the seismogenic ruptures is theoretically possible, however known strong historical earthquakes in the northern Tien Shan demonstrate maximum length of the rupture barely reaching 200 km: for example, Kebin (Kemin) earthquake of 1911 (Bogdanovich et al., 1914). It is possible that in such (and probably in some other) cases it took place an artificial unification of different earthquakes occurred in different parts of the fault, but in close time frames. Nevertheless one cannot exclude a possibility of propagation of many segments along almost whole plane of the Talas-Fergana Fault (during both discussed earthquakes there were united 11 segments in three chains of the disjunctive). An example of such propagation in the Tien Shan is mentioned Kebin earthquake during which there was a propagation of 6 fault segments of the fault zone close in time (Delvaux et al., 2001).

A distribution of paleoearthquakes along the Talas-Fergana Fault in time (for exclusion of individual "jumps") has clear evidence (Fig. 42). In interval 6000-4500 years ago the strong earthquakes occurred in north-western chain of the fault. Then a seismic activity has spread over to the south-eastern chain: 4500-2500 years ago. In the interval 2500-1500 years ago the strong earthquakes occurred in the central and south-eastern chains of the fault. Then the seismic activity is concentrated in north-western and central chains: 1500-250 years ago.

1350 ±60 1450 ±40 2020 ±50 2020 ±50

Sulu-Bakair 5210 ± 155 5210 ± 155

materials on the Talas-Fergana Fault from Kazakhstan and China.

fault segments of the fault zone close in time (Delvaux et al., 2001).

Table 4. Radiocarbon dates of samples collected from displaced gullies along the Talas-Fergana Fault (by data of Korjenkov et al., 2009, 2010, as well as V.S. Burtman et al., 1987,

chain starts from Kyldau river valley and ends in a region of southern state border of Kyrgyz Republic. We understand whole conventionality of our division, especially without

Because of uneven investigations and bareness of existed materials one has to talk very carefully on segmentation of the Talas-Fergana Fault zone. According to existing data we select 13 segments: 3 – in the northwestern chain of the fault, 5 or 6 in the central chain and 4

An analysis and comparison of materials of Table 4 and Fig. 41 have allowed us to reveal 18 paleoseismic events, 17 of which occurred in the second half of Holocene (Table 5 and Fig. 42). We assessed also distances between localities, where there were determined absolute ages of the seismogenic displacements, occurred (supposedly) during one seismic event. We conditionally accepted these distances as minimum lengths of the

Some of the extreme values of rupture lengths, such as 270 km and 220 km for earthquakes occurred in 4530 and 1980 years ago, provoke a natural doubt. Although such length of the seismogenic ruptures is theoretically possible, however known strong historical earthquakes in the northern Tien Shan demonstrate maximum length of the rupture barely reaching 200 km: for example, Kebin (Kemin) earthquake of 1911 (Bogdanovich et al., 1914). It is possible that in such (and probably in some other) cases it took place an artificial unification of different earthquakes occurred in different parts of the fault, but in close time frames. Nevertheless one cannot exclude a possibility of propagation of many segments along almost whole plane of the Talas-Fergana Fault (during both discussed earthquakes there were united 11 segments in three chains of the disjunctive). An example of such propagation in the Tien Shan is mentioned Kebin earthquake during which there was a propagation of 6

A distribution of paleoearthquakes along the Talas-Fergana Fault in time (for exclusion of individual "jumps") has clear evidence (Fig. 42). In interval 6000-4500 years ago the strong earthquakes occurred in north-western chain of the fault. Then a seismic activity has spread over to the south-eastern chain: 4500-2500 years ago. In the interval 2500-1500 years ago the strong earthquakes occurred in the central and south-eastern chains of the fault. Then the seismic activity is concentrated in north-western and central chains: 1500-250 years ago.

Kara-Bura

1996 and Trifonov et al., 1990, 1992)

or 5 in southeastern chain (See Fig. 41).

seismogenic ruptures.

How the fault will behave in future is not clear. Here apparently it can be two variants of events' development: 1) the seismic activity will continue in north-western and central chains of the fault or 2) most probably will spread over to its south-eastern chain, where now there is so-called "a seismic gap": last earthquake occurred here already 1720 year ago!

Fig. 41. Distribution of the earthquakes and segmentation along the zone of the Talas-Fergana fault in bounds of the Kyrgyz Republic

Fig. 42. Migration of the earthquakes along the Talas-Fergana Fault zone in Holocene


Fig. 42. Migration of the earthquakes along the Talas-Fergana Fault zone in Holocene


Table 5. Average (calculated) ages of strong earthquakes occurred in the Talas-Fergana Fault zone in Holocene, interval between them and minimum length of the seismogenic rupture.

An important question in a concern of the long-term forecast of strong earthquakes is their reoccurrence. We have calculated intervals between strong earthquakes along the TalasFergana Fault during the second half of Holocene (Table 5) where representation of the earthquakes is more complete. They rank from 145 to 850 years. Thus, an average calculated reoccurrence of the earthquakes along the whole zone of the Talas-Fergana Fault is 375 years. However, an arithmetic mean value is not the best characteristics of the natural phenomena. Comparison of number of the strong earthquakes along the fault with interval of their occurrence (Fig. 43) has allowed us to reveal three clear peaks of the earthquake occurrence in the second half of Holocene divided by intervals in 300 years.

Fig. 43. Comparison of number of the strong earthquakes along the Talas-fergana Fault with intervals of their occurrence

Taking into account all said above, we can supposed that the next strong earthquake (M>7) most probably will occur in approximately 25 years: 300 years minus 275 years (an age of the last strong paleoearthquake) in south-eastern limb of the Talas-Fergana Fault.

We have discussed above that we have conducted a segmentation of the Talas-Fergana Fualt zone by existed data (Fig. 41). In result there were revealed 13 segments: from "a" to "m". We wrote also that by previous studies (Delvaux et al., 2001) there was revealed that during strong Tien Shan earthquakes it takes place a unification of several segments of the fault zone. We pointed above that during Kemin earthquakes of 1911 (M=8.2) it took place a unification of 6 segments of the Chilik-Kemin seismogenic zone of total length to 200 km (Delvaux et al., 2001). The same unification of the segments could take place also during strong earthquakes which occurred along the Talas-Fergana Fault zone (Table 6).


Fergana Fault during the second half of Holocene (Table 5) where representation of the earthquakes is more complete. They rank from 145 to 850 years. Thus, an average calculated reoccurrence of the earthquakes along the whole zone of the Talas-Fergana Fault is 375 years. However, an arithmetic mean value is not the best characteristics of the natural phenomena. Comparison of number of the strong earthquakes along the fault with interval of their occurrence (Fig. 43) has allowed us to reveal three clear peaks of the earthquake

Fig. 43. Comparison of number of the strong earthquakes along the Talas-fergana Fault with

Taking into account all said above, we can supposed that the next strong earthquake (M>7) most probably will occur in approximately 25 years: 300 years minus 275 years (an age of

We have discussed above that we have conducted a segmentation of the Talas-Fergana Fualt zone by existed data (Fig. 41). In result there were revealed 13 segments: from "a" to "m". We wrote also that by previous studies (Delvaux et al., 2001) there was revealed that during strong Tien Shan earthquakes it takes place a unification of several segments of the fault zone. We pointed above that during Kemin earthquakes of 1911 (M=8.2) it took place a unification of 6 segments of the Chilik-Kemin seismogenic zone of total length to 200 km (Delvaux et al., 2001). The same unification of the segments could take place also during

the last strong paleoearthquake) in south-eastern limb of the Talas-Fergana Fault.

strong earthquakes which occurred along the Talas-Fergana Fault zone (Table 6).

intervals of their occurrence

occurrence in the second half of Holocene divided by intervals in 300 years.


Table 6. Lengths of the seismogenic ruptures along the Talas-Fergana Fault zone and possible magnitudes of earthquakes.

We have analyzed number of formulae for determination of paleoseismic catastrophes' magnitudes according to parameters of seismic rupture, published by different investigators: V.P. Solonenko and V.S. Khromovskikh (1978), A.A. Nikonov (1984) and D.L. Wells and K. J. Coppersmith (1994). Let's investigate their formulae for magnitude assessment by a length of the seismogenic rupture expressed in the surface and check these results on measured parameters of the fault scarps and an instrumental magnitude of the Tien Shan's Suusamyr earthquake of 1992 occurred in the depression with the same name (Bogachkin et al., 1997).

During the earthquake in the surface there were occurred only two short seismogenic ruptures with a total length of 4 km, a distance between then was 26 km (Bogachkin et al., 1997). A magnitude assessed instrumentally was Мs = 7.3. As it was discussed above at a description of the Suusamyr earthquake, in this case we have a deal with so-called "blind" seismogenic rupture, larger part of which did reach the surface. Let's assume that the total length of the rupture (L) was 4 + 26 km = 30 km.


M = 0.6 lg L + 6, we got М = 6.89.


 $\mathbf{M} = \text{6.61 + 0.55} \text{ lg L}$   $\text{we get M} = 7.42...$ 


$$\text{M} = 5.08 + 1.16 \text{ lg L} \text{, we get } \text{M} = 6.79.$$

Cited above calculations show that the data by V.P. Solonenko and V.S. Khromovskikh (1978) on earthquakes parameters of the Baykal Lake and Caucasus regions, as well as world data by D.L. Wells and K.J. Coppersmith (1994) give underestimated magnitude of the earthquake if compare with the instrumental value. At the same time the formula by A.A. Nikonov (1984), calculated by him for earthquakes of the Central Asia gives a value which only on 0.1 higher than an instrumental value. This is a very good result especially if we are taking into an account that an accuracy of magnitudes determination by such method is in bounds of ±0,5 of the magnitude unit. This is why in our magnitudes assessments we based on formula by A.A. Nikonov (1984), which he specially deduced for a territory of the central Asia.

Our calculation (Table 6) have shown that according to paleoseismological data along the Talas-Fergana Fault zone there are possible earthquakes with magnitude М>7, and during unification of many segments (up to 11) a maximum magnitude can reach М=8. One can not exclude however that along the fault zone there were occurred two or more independent earthquakes divided by short time intervals. This interval we can not reveal because of significant miscalculations of Radiocarbon method of dating. It is possible that there were a clustering of the earthquakes along the seismogenic zone. In a history of strong earthquakes of the Tien Shan such clustering took place in the end of XIX – beginning of XX centuries. Here along so-called Northern Tien Shan Seismic Zone during only 26 years there were occurred 4 strong earthquakes: Belovodsk one of 1885 with МLH = 6,9; Verny earthquake of 1887 with МLH = 7,3; Chilik one of 1889 with МLH = 8,3 and Kebin earthquake of 1911 with МLH = 8,2 (Dzhanuzakov et al., 2003). If paleoseismologists will study consequences of those earthquakes in 3011 using the Radiocarbon method, then because of miscalculations of the method, for them it will be that different segments of the Northern Tien Shan Seismogenic Zone activated simultaneously in 1900 AD plus-minus 50 years…

Thus we understand all conditionality and approximateness of cited above attempts to conduct of the segmentation of the Talas-Fergana Fault zone and calculations of magnitudes of paleoearthquakes with use of so scanty data along the fault zone of 350 km length only in Kyrgyzstan territory. However one has to start from something. The future materials on age of displaced of relief elements, full-fledged paleoseismological trenches, which will cross the whole fault zone, will help to define more precisely cited above numbers.
