1. Introduction
Landslides have been the focus of many studies in recent years due to their disruptive and destructive effects on human life and property. Expansion of urban centers through construction of roads and buildings has had a destabilizing effect on natural soil slopes. Though this effect has been increasing in recent time, there are other factors that cause destabilization in soils that are produced by the environment. Soil properties like clay content and permeability have a considerable effect on a soils ability to retain its structural strength under substantial loads or rainfall [1] [2]. It is common knowledge that rain affects the stability of soil in terms of load- bearing capacity due to the fact that as rain infiltrates into the soil, the soil moisture increases, which reduces the cohesion and shear strength of the soil [3][4].
The established relationship between rain and soil stability has led researchers, such as [5][6][7], to be more and more concerned with how much rain in given conditions will trigger or influence a landslide or other mass movements. Recently, statistical analysis of historical rain data in conjunction with historical landslide occurrences has enabled researchers to begin the process of creating modelling or prediction tools to characterize rain events as potential landslide-triggering events [7][8][9]. Several studies deal with the use of rainfall thresholds as a method to sort out any landslides that were not caused by rainfall to isolate the rainfall amounts that caused landslides in the past [7][8][10]. These tools would then allow governments to use weather forecasts to predict the possibility of landslides and mobilize the necessary resources to reduce the amount of damage caused by the event. Although these methods of prediction do not necessarily give specific locations where the landslides would occur, they do provide warning of precipitation events that could cause failure in any unstable slopes. The purpose of this study is to analyze damage caused by landlsides and the magnitude of the cumulative rainfall preceding the landslide. This would then identify areas that are more prone to landslides than others based on historical data.
Panama is a country that experiences landslides frequently.
Panama has a tropical climate that experiences only dry and rainy seasons that are characterized by little rain and large, intense rain events respectively [11]. The combination of the tropical climate and the increase in construction throughout urbanizing areas has resulted in larger numbers of landslides and greater subsequent impacts.
Panama’s soil types also have considerable effects on landslide occurrences due to most of the soil types having considerable clay contents. Soils like Andosols and Nitisols are common (figure 1), but there is considerable variation in soil type in boundary areas where geology changes.
Figure 1. General soil map of Panama. Legend. Ah: Humic Andosols, Nd: dystric Nitisols, Od: dystric Histosols. Source: Food and Agriculture Organization (FAO) and United Nations Educational, Scientific, and Cultural Organization (UNESCO) [12].
The purpose of this study is to determine what districts in the province of Chiriquí is most susceptible to landslides and what the rainfall conditions are in those districts.
2. Methodology
2.1 Study Area
The province of Chiriquí is on the western side of Panama and shares a border with Costa Rica. It is 6,491 km 2 (2506.2 mi2) in area and constitutes the highest portion of the western highlands with a maximum elevation of 3.48 km (2.16 mi.) above sea level. The province is split up into 14 districts: Alanje, Barú, Boquerón, Boquete, Bugaba, David, Dolega, Gualaca, Remedios, Renacimiento, San Félix, San Lorenzo, Tierras Altas, and Tolé. It was chosen as the area of study due to the abundance of meterological stations and thereby meteorological data for that region. Landslide occurrences are becoming more common in this province, as there have been 22 significant landslide events between september and october, from 1986 to 1998.
The soil type of most of the region has played a part in this increase. Some of the soil types present in Chiriquí are a direct result of the weathering of underlying volcanic geology from Volcano Baru. The best example of this is represented by the ANum symbol in figure 2, which stands for Andosol. This type of soil is relatively young, has mineralogical similarities with its parent material, and low clay content.
Figure 2. Soil map of Western Panama. (Legend:LPdy: Dystric Leptosols, CMfl: Ferralic Cambisols, ACct: Cutanic Acrisols, AChd: Haplic Acrisols, ACpl: Plinthic Acrisols, NTum: Umbric Nitisols, ANum: umbric Andosols) Source: Atlas de suelos de America Latina y el Caribe. (2014) [13].
The majority of Chiriquí is dominated by varieties of Acrisols (figure 2), which are often susceptible to erosion and dominated by stable, low-activity clays. There are also other soils in Chiriqui, like Andosols and Leptosols, which have low clay content, but are still influenced by rain to produce landslides once the soil is saturated enough to decrease surface friction (Seed, et al., 2008). Overall, rain is a major factor in both the formation of the soils present in Chiriquí as well as on the soils structural stability.
2.2 Data collection
Rain data was collected from the online meterological database operated by Empresa de Transmisión Eléctrica, S.A.(ETESA), which is the governmental agency responsible for the development and maintenance of the power lines in Panama. The data collected were daily measurements of precipitation taken from 23 stations across the entirety of Chiriquí. The time frame of the data was from May through December and reaching from 1986 to 2010. The purpose for choosing May through december was to analyze the average amount of rain each district received over the course of the rainy season to establish average precipitation conditions in Chiriquí. The measurements from these stations were used for statistical analysis of the precipitation on each day within the study period (june-december 1986-2010). The data for the landslides were downloaded from a database called DesInventar, which is operated and supported by Sistema Nacional de Proteccion Civil (SINAPROC) and United Nations International System for Disaster Reduction (UNISDR), among others. The database contains records of disaster events which contain information on what elements were affected (i.e. number of people affected, number of deaths, number of homes affected) and the event that was responsible, like landslides. This data was compared to analyze for trends.
2.3 Statistical analysis
The daily rainfall averages were used to produce yearly average rain amounts received by each district. The data was graphed in a scatter plot and then compared to the landslide data generated by the DesInventar database. Many other studies use more complicated statistical and modelling methods to characterize rain events using much more specific parameters [3][6][7][8]. Comparison of the overall amount of precipitation received in each district with the landslides that occurred within the study period there provided insight into which districts in Chiriquí are more at risk to landslides.
3. Results
Boquete and Bugaba districts were found to have experienced the most landslides in Chiriquí during the study period. This can be seen in figure 3. Bugaba was by far the most prone to landslides with a total of 15 landslides occurring within the study period. Boquete followed in landslide frequency with 8 landslides occurring within the study period. Tole, Baru, and David experienced equal numbers of landslides, but only fractions of the number of landslides that occurred in Boquete and Bugaba.
Figure 3. Landslide frequency for each district over the study period.
The rainfall data collected and averaged is displayed in figure 4 and gives further insight into the average precipitation conditions for each of the districts in question. Overall, San Felix seems to receive the greatest daily rainfall depths, but experienced the least landslides during the study period. Boquete and Bugaba received similar rainfall depths to districts that did not experience nearly as many landslides.
Figure 4. Average Daily Rainfall depths over the study period for each district in Chiriquí affected by landslides in the study period.
4. Discussion
The data presented in figures 4 and 5 seem to indicate that Boquete and Bugaba are the most prone to landslides in Chiriquí. The landslides that occurred in Boquete and Bugaba were possibly caused by lower average rainfall depths than other districts with fewer landslides. This is most likely due to the fact that Bugaba and Boquete have young, reactive soil that is more unstable than the more stable soil in areas like Tole and San Felix (see figure 2). Tole, Baru, and David are the next most susceptible to landslides and less if not more rainfall on average.
5. Conclusions