Preliminary seismic hazard of Central America

Our estimations are based on a seimotectonic zonation of the Central America region, and in an improved regional earthquake catalog recently obtained. Through a detail research of different earthquake catalogue sources and archives, the new catalog were compiled. In order to have magnitude standarization and homogeneity, we convert previously obtained magnitudes (Ms, mb and ML) to a moment magnitude Mw scale. We then applied a strong motion attenuation relationship developed from Central American strong motion data.

2- INTRODUCTION 2

3- METHODOLOGY 2

4- THEORETICAL BACKGROUND 3
4.1 Strong motion attenuation relationship 4

5- HISTORICAL SEISMICITY 4

6- SEISMOTECTONIC MODELING 5
6.1 Middle America Subduction Zone 6
6.2 Middle America Volcanic Arc 6
6.3 Polochic-Motagua Fault System 7
6.4 Guayape Fault System 8
6.5 Honduran Depression Fault System 8
6.6 Median Trough Faults and Nicaraguan Depression 8
6.7 The intraplate Caribbean Sea 8
6.8 North Panama Deformed Belt 9
6.8.1 West Panama an eastern Costa Rica Deformed Belt 9
6.8.2 North Central Panama Deformed Belt 9
6.8.3 Northeastern Panama Deformed Belt 9
6.9 The Panama Fracture Zone 10
6.10 Southwestern Panama 10
6.11 Central southern Panama . . . 10
6.12 The Atrato Suture Zone 10
6.13 Intermediate depth Seismic Zones 10
6.13.1 Source areas with depths from 40 to 70km 11
6.13.2 Source areas with depths from 70 to 110km 11
6.13.3 Source areas with depths from 110 to 200km 12

7- SEISMOLOGICAL DATA BASE 12
7..1 The Catalog 12
7.2 Aftershock cleaning 12
7.3 Magnitude standarization 12
7.4 Completeness Analysis 13

8- PARAMETERIZACION OF SEISMIC SOURCES 13
8.1 Estimation of the Maximum Magnitude 13
8.2 Recurrence values: a and b 14
8.3 N-values 14

9- RESULTS OF THE SEISMIC HAZARD ASSESSMENT OF THE CENTRAL AMERICA REGION 14

ACKNOWLEDGEMENTS 15

REFERENCES 17

APENDIX 20

The earthquake hazard assessment results we here obtain, represent a first order approach to the actual problem. We have employed parameter variability, using a logic tree approach, which will then allow us to compute confidence intervals for the estimates of PGA (peak ground aceleration).

The logic tree formalism were performed following this criteria: all parameters were given with two extreme values of b, they were taken as the center value plus and minus 0.15, with weights of 0.2, 0.6 and 0.2. The extreme values of N, were taken as the center value divided and multiplicated by 2, with weigths 0.2, 0.6 and 0.2. The maximum magnitude extremes were taken to be the center value minus 0.2 to 0.5 (depending of the seismic source) and plus 0.2, with weights 0.2, 0.6 and 0.2. The depths were weighted for most of the zones, as 0.25, 0.5 and 0.25, where the extremes were taken as the outer volume bounds. An example of expected PGA (m/s2) grafic hazard for an specyfic site is shown in Fig. 37, and an. example of results of the hazard analysis for an especific site.

The PGA curves of the seismic hazard maps (Figs. 10 to 13) seem to be controlled by the seismicity related to the subduction, and in much minor degree by the shallow fault systems. That may be not represent the reality specially for some zones, for example: along the volcanic arc or in the case of Polochic-Motagua seismic zone that are not adecuatelly modelled.

The hazard curves values tend to increase from, the northwest to the southeast along the pacific margin of Central America, achieving greater PGA values in Nicaragua and Costa Rica. This could be related to the increment in the Cocos-Caribbean interplate convergence rate, that varies from 7.0 cm/yr in front of Guatemala to 9.5 cm/yr at the Costa Rica margin. Another possibility is that the PGA increment is an calculation artifact due to an inadequate estimation of the maximum magnitudes and also of the recurrence intervals for the characteristic subduction earthquakes indifferent sectors along the interplate interphase.

In general, the isoaccelerations curves seems to be low according to our spectations, specially along some areas of the inner arc region (for example, in San Salvador, Managua and in the Central Valley region of Costa Rica). But this is to be spected, because at a regional level is not possible to model appropriately some areas with a spected higher hazard. The curves we show are very smoothed and controlled by the more regional tectonic structures. We have to understand also, that they are for a rock site conditions, and we did not have made this regional calculations for soil level, because there is a lot of variability of soil conditions along the region.

The final point we would like to emphasize here, is that the present model is a preliminary regional seismic hazard zonation, therefore, it is not really appropiate to use in some local sites, where there exists much highly relevant information about local faulting, which is smoothed in a regional modelling.