Hillslopes in humid‐tropical climates aren't always wet: Implications for hydrologic response and landslide initiation in Puerto Rico

Abstract

The devastating impacts of the widespread flooding and landsliding in Puerto Rico following the September 2017 landfall of Hurricane Maria highlight the increasingly extreme atmospheric disturbances and enhanced hazard potential in mountainous humid-tropical climate zones. Long-standing conceptual models for hydrologically driven hazards in Puerto Rico posit that hillslope soils remain wet throughout the year, and therefore, that antecedent soil wetness imposes a negligible effect on hazard potential. Our post-Maria in situ hillslope hydrologic observations, however, indicate that while some slopes remain wet throughout the year, others exhibit appreciable seasonal and intra-storm subsurface drainage. Therefore, we evaluated the performance of hydro-meteorological (soil wetness and rainfall) versus intensityduration (rainfall only) hillslope hydrologic response thresholds that identify the onset of positive pore-water pressure, a predisposing factor for widespread slope instability in this region. Our analyses also consider the role of soil-water storage and infiltration rates on runoff generation, which are relevant factors for flooding hazards. We found that the hydro-meteorological thresholds outperformed intensity-duration thresholds for a seasonally wet, coarse-grained soil, although they did not outperform intensityduration thresholds for a perennially wet, fine-grained soil. These end-member soils types may also produce radically different stormflow responses, with subsurface flow being more common for the coarse-grained soils underlain by intrusive rocks versus infiltration excess and/or saturation excess for the fine-grained soils underlain by volcaniclastic rocks. We conclude that variability in soil-hydraulic properties, as opposed to climate zone, is the dominant factor that controls runoff generation mechanisms and modulates the relative importance of antecedent soil wetness for our hillslope hydrologic response thresholds.

Comparing methods for determining landslide early warning thresholds: potential use of non-triggering rainfall for locations with scarce landslide data availability

“…The different thresholds for the Frequentist N method vary little and are not visible in the plot must be carefully checked for errors and thus possibly removed. If they are not errors in the dataset, the presence of low-intensity triggering points is a sign that only intensity and duration may not be sufficient to obtain a threshold with the necessary reliability, and that approaches taking into account predisposing hydrological factors must be investigated Greco 2016, 2018;Thomas et al 2019Thomas et al , 2020Marino et al 2020). As an overall conclusion, the analysis we have carried out suggests that future studies for threshold determination for newly investigated areas may start using the Frequentist N method first, and then move to the Frequentist PN as more data becomes available, as this method delivers globally the best of the techniques explored herein.…”

A Catalogue of Tropical Cyclone Induced Instantaneous Peak Flows Recorded in Puerto Rico and a Comparison with the World’s Maxima

“…Even though the largest amounts of TC rainfall have typically occurred in the eastern, southeastern, and central interior portions of the island [34,75], rainfall intensity and totals are largely dependent on the specific interactions of internal TC moisture, and TC wind direction with topographical features [58]. Runoff generation in PR is preferentially controlled by subsurface stormflow and saturation overland flow mechanisms for forested areas due to the high infiltration rates [76] and abundance of macropores [77] however, it is prone to the occurrence of excess precipitation overland flow on disturbed areas [78]. However, studies conducted elsewhere show that dramatic changes in land cover are typically undetectable by empirical flow data even in small watersheds (e.g., 35 ha watershed under controlled experiment conditions described by Birkinshaw, et al [79]) and its effects dramatically diminish with both increasing drainage area and storm size [80].…”

Simulated or measured soil moisture: Which one is adding more value to regional landslide early warning?

“…Simplification and misrepresentation of some of these processes in the CoupModel set-up may lead to an underestimation of the seasonal soil 490 moisture variation, which in Switzerland is high, with generally wet conditions from fall to spring and a dry period with intermittened wetting events during summer (Pellet and Hauck, 2017). A limited seasonal representation may reduce the forecast model's ability to separate triggering from non-triggering conditions, as reported for regional landslide forecast models where regions with different seasonal soil moisture variation were compared (Thomas et al, 2020). Particularly in regions with a high seasonal evapotranspiration variation, wet and dry periods may be controlled by different soil moisture fluxes with 495 vertical fluxes being dominant during dry periods and lateral fluxes during wet periods (e.g.…”

“…While this approach benefits 45 from widely available rainfall data, the focus on triggering factors disregards the influence of the antecedent wetness conditions ("cause factors"), which could be represented by including soil wetness information (Bogaard and Greco, 2018). In fact, forecast goodness improvement was reported after incorporation of in situ soil moisture measurements (Mirus et al, 2018a(Mirus et al, , 2018bThomas et al, 2020), remotely sensed soil moisture (Bordoni et al, 2020;Brocca et al, 2016;Thomas et al, 2019;Zhao et al, 2019a;Zhuo et al, 2019b) or simulated soil moisture using physically-based models (e.g. Ponziani et al, 2012;50 Segoni et al, 2018b;Zhuo et al, 2019a).…”