The potential impact of climate change on typhoon-triggered landslides in Taiwan, 2010–2099

Abstract

In the western North Pacific, including Taiwan, typhoons (tropical cyclones) and rainfall rates are predicted to intensify as a result of climate change. Because typhoons are the major triggers of shallow landslides in Taiwan, landslide activity is expected to increase as global warming continues. To assess the worst scenario of landslide occurrence in a mountainous watershed till the end of the century, this study developed a method to select a global climate model (GCM) from 21 available GCMs and correct its monthly precipitation data, before downscaling annual maximum (24-h) rainfall from the corrected GCM data as input to the factor-of-safety model for landslide prediction. Average annual maximum rainfall is expected to increase from 322 mm in 1960–2008 to 371 mm in 2010–2099. Average total unstable area is expected to increase from 1135 ha in 1960–2008 to 1280 ha in 2010–2099, a 12% increase. As a first attempt to assess landslide activity due to global warming, this study is useful as a reference for watershed management in Taiwan. The results must be evaluated in light of uncertainties caused by the correction and downscaling of GCM data and the input parameters to the slope stability model.

Introduction

Global warming is unequivocal according to the latest summary report of the Intergovernmental Panel on Climate Change (IPCC, 2007). As sea surface temperature (SST) increases, water vapor in the lower troposphere also increases. Both conditions tend to increase the energy for the development of tropical cyclones (typhoons). The relationship between global warming and increased typhoon activity and intensity has been confirmed by studies based on analyses of historical records and simulation results (Knutson and Tuleya, 2004, Emanuel, 2005, Trenberth, 2005, Webster et al., 2005, Hoyos et al., 2006, Allan and Soden, 2008, Elsner et al., 2008).

As part of the global pattern, typhoons in the western North Pacific (WNP), including Taiwan, have also become more intense as measured by their frequency in hurricane categories 4 and 5 (Webster et al., 2005) or the power dissipation index (Emanuel, 2005). The mechanism linking global warming to typhoon activity in the WNP, however, is more complicated. Some studies have traced the linkage to the El Niño–Southern Oscillation (ENSO), rather than local SSTs: in El Niño years, when above-normal SST appears over the equatorial eastern Pacific Ocean, typhoons tend to be more intense and have longer lifetimes (Chan and Liu, 2004, Camargo and Sobel, 2005, Stowasser et al., 2007, Chou et al., 2009). Still others have claimed that typhoon activity in the WNP is influenced by changes of typhoon formation locations and prevailing tracks due to global warming (Wu and Wang, 2004, Wu and Wang, 2008, Tu et al., in press). Regardless of its mechanism, climate change has caused, and is likely to cause, more intense typhoons in the WNP.

Typhoons are major triggers of shallow landslides in mountainous watersheds of Taiwan (21–26°N and 119–125°E). In this study we hypothesize that the predicted intensification of typhoon activity will have severe consequence on landslide occurrence. But, how severe can it be? This is the basic question for the study. It is also an important question for watershed management and disaster reduction in Taiwan. The literature relating climate change to landslide activity is limited. And, to our best knowledge, no attempt has been made to assess landslide activity in a climate change scenario involving extreme rainfall events such as typhoons. Therefore, can current global climate models (GCMs) provide adequate data for assessing typhoon-triggered landslides in a climate change scenario? This is another basic question addressed in this study.

In 1994–1996 the European Commission sponsored a project (“The temporal stability and activity of landslides in Europe with respect to climatic change,” or TESLEC) examining the effect of global warming on landslide activity at selected sites in Europe (Dikau and Schrott, 1999). TESLEC included studies that used general circulation models to downscale local climate change scenarios, before using them as input for predicting landslide activity in a changing climate (Buma and Dehn, 1998, Dehn and Buma, 1999, Collison et al., 2000, Dehn et al., 2000). These studies showed the feasibility of linking GCMs and a slope stability model via downscaling techniques, but they also highlighted the uncertainties associated with the approach and the importance of other factors such as changing slope geometry and vegetation succession that might also affect landslide activity.

More recently, Dixon and Brook (2007) applied climatic changes predicted by the GCM UKCIP 2002 to a rainfall threshold model and reported that the return period for the reactivation of an active landslide in Derbyshire, U.K. could decrease from 4 to 3.5 years by the 2080s. In their study of the southwest coast of British Columbia, Canada, Jakob and Lambert (2009) predicted a 28% increase in the total number of debris flows by the end of the century due to increases in antecedent precipitation and 24-h precipitation intensity as predicted by an ensemble of 19 GCMs. And Jomelli et al. (2009) expected a reduction of the occurrence of hill slope debris flows in the French Alps by the end of the century due to a predicted decrease in intense rainy events and an increase in temperature using the GCM ARPEGE.

Like previous studies, this study also uses a downscaling approach to derive precipitation data from a GCM as input to a watershed-level landslide prediction model. Because this study focuses on the impact of typhoons on landslide activity, extra steps in data processing are needed to extract extreme rainfall data. Inevitably, the process of downscaling and approximation introduces uncertainties into the results. Therefore, we approach this climate impact study from the perspective of susceptibility analysis rather than as a prediction problem (Johnson and Weaver, 2009). Our principal objective is to assess the worst scenario of landslide occurrence in a mountainous watershed in Taiwan during the 21^st century due to intensified typhoon activity.