16HEALTH Modeling Landslides Associated to Heavy Rainfall in an Urban Area Through an Integrated Risk Approach

The risk of landslides in urban areas is increasingly frequent. It is estimated that worldwide, 80% of them, are induced by extreme rainfall (Froude, M. J. and Petley, D. N., 2018) (Ozturk, U. et al., 2022), so that, according to climate change forecasts which consider an increase in heavy rainfall, the risk of landslides will increase. However, since a risk perspective, other factors underlying the complex problem of landslides in urban areas are decisive.

Currently, landslides in urban areas are beginning to be relevant in disaster risk research, given the amount of population and property exposed, and the ever-increasing urbanization phenomenon.

The present study analyzes landslide susceptibility factors (vulnerability), as well as changes in intense precipitation (hazard or threat), which are involved in the occurrence of landslides in the city of Xalapa, Veracruz. In addition, since the characteristics of the territory are dynamic, a two-period comparison was carried out to determine the changes in risk. For the analysis of precipitation, statistical calculations were made to identify variations over time in the precipitation pattern.

The precipitation can vary in the study area due to both anthropic influences and natural variability (Vargas, N. and F. Rodríguez, 2024) (Ma, F. and Yuan, X., 2021), landscape modifications due to urbanization, among other changes in land use, can be as important factors as the presence of intense rainfall. Therefore, it is convenient to analyze quantitatively the factors that induce the hazard (heavy rains), as well as the factors that induce vulnerability (susceptibility to landslides and socioeconomic conditions), and thus determine the most appropriate measures to make disaster risk management more accurate.

Although it is well known that precise weather and climate data are scarce in middle and low income countries such as Mexico, we worked with various sources of data, such as those from the National Meteorological Service, available at https://sih.conagua.gob.mx , data from the Municipal Secretariat of Civil Protection of Xalapa, and reanalysis data for precipitation, available at: https://app.climateengine.org/climateEngine.

In the case of landslide disasters, various sources of information were consulted, such as: EM-DAT, DesInventar and those provided personally by the Civil Protection Secretariat of the municipality of Xalapa, Veracruz.

Although in both cases there are inconsistencies and given the lack of a landslide inventory in the study area, it is emphasized that the risk analysis process is an approximation based on available data.

Method

We worked with layers of information in a GIS to obtain landslide susceptibility in the city of Xalapa, Veracruz, for the years 1995 and 2020 (available data), considering the following factors: soil type, land use and vegetation type, geomorphology, slope, proximity to watercourses and socioeconomic conditions of the exposed population. Secondly, precipitation data were obtained for the area for the two study periods, and were subjected to statistical analysis to determine changes in the precipitation pattern, such as the annual and monthly accumulated amount, changes in precipitation intensity on 20-year scales, among other indicators that lead us to show differences in the behavior of rainfall over time.

Results

Through the analysis we found both changes in precipitation behavior and changes in the landscape. In other words, for both periods the contexts are different both in terms of vulnerability and hazard or threat conditions, the latter represented by rainfall. The changes detected in the rainfall pattern may have different origins, both anthropogenic and natural variation. In the case of vulnerability or susceptibility, the changes detected are specifically of anthropic origin, which represents an opportunity to modify the course of events.

Finally, in the model validation stage, we found that our results coincided with the location of landslides, specifically for the 2020 period, since for the previous period the data we have are less accurate.

Conclusions

In the current context, there are changes at various levels that should be considered for landslide assessments and analysis. From a risk approach, changes in climate and meteorological weather can be observed; in the case of landslides in which rainfall is an important trigger, different precipitation patterns can be observed, even on a decadal scale. On the other hand, in the vulnerability factors (susceptibility), changes such as land use can occur, which represent a great weight in the quantification of the factors involved in the occurrence of landslides. However, in early warning systems, the focus of the analysis is on determining precipitation thresholds.