The link between rainfall intensity and global temperature

The aftermath of a landslide in Taiwan caused by very heavy rainfall

One of the most interesting aspects of the global landslide database that we maintain at Durham is the way in which it has highlighted the importance of rainfall intensity in the triggering of fatal landslides. Generally speaking, to kill people a landslide needs to move quickly rapid, and rapid landslides appear to be primarily (but note not always) triggered by intense rainfall events (indeed in the reports the term "cloudburst" often crops up). So, a key component of trying to understand the impacts of human-induced global climate change on landslides is the likely nature of changes in rainfall intensity, rather than that of rainfall total. Put another way, it is possible that the average annual rainfall for an area might decrease but the occurrence of landslides increase if the rainfall arrives in more intense bursts.

There is of course a certain intuitive logic in the idea that rainfall intensity might increase with temperature. Warmer air is able to hold more moisture (as anyone who has been in the subtropics in the summer will know only too well!) and of course increased temperatures also drive greater convection, responsible for thunderstorm rainfall. Of course this is a very simplistic way to look at a highly complex system, so it is not enough to rely upon this chain of logical thought. However, until now there have been surprisingly few studies to actually quantify whether there is a relationship between global temperature and precipitation intensity, which has meant that for landslides understanding the likely impact of climate change has been quite difficult.

However, an important and rather useful paper examining exactly this issue has sneaked under the radar in the last few months. The paper, by Liu et al (2009) (see reference below), was published in Geophysical Research Letters a couple of months ago. The paper uses data from the Global Precipitation Climatology Project (GPCP). These data can be accessed online here (so no claims that climate scientists don't publish their data, please!) The dataset provides daily rainfall totals for 2.5 x 2.5 degree grid squares across the globe, extending back almost 50 years. Liu et al. (2009) looked at the data from 1979 to 2007, comparing precipitation density with global temperature in this time period.

Their results are both unsurprising and surprising. The unsurprising part is that they found that the occurrence of the most intense precipitation events does increase with temperature. The surprising part is the magnitude of the change - they found that a 1 degree Kelvin (Centigrade) increase in global temperature causes a 94% increase in the most intense rainfall events, with a decrease in the moderate to light rainfall events. Indeed the median rainfall increased from 4.3 mm day⁻¹ to 18 mm day⁻¹, which is a surprisingly high shift as well.

So why is this important in the context of landslides? Well, I think that there are probably two key implications:

1. It has long been speculated that anthropogenic warming will lead to an increase in landslides, but with little real quantitative evidence to confirm or deny this. The demonstration that higher global temperatures does lead to increased precipitation intensity starts to put some meat on the bones of this idea. Furthermore, if it is possible to directly link rainfall intensity to landslide occurrence (and there is some evidence both from my own work and from that of others that this may be possible), then it should be possible to start to examine the likely increase in landslides as warming proceeds.
2. The current global climate models assume a much lower increase overall in precipitation intensity with increasing temperature than Liu et al. (2009) suggest. Indeed most of the models assume about a 7% increase per degree Kelvin (Centigrade) warming. For the most intense precipitation events this means that the models predict about a 9% increase, which is an order of magnitude lower Liu et al. (2009) found. This suggests that the rainfall projections that are derived from the models are probably overly-conservative, and possibly very much so, which is a concern. If so, then forecasts of landslide occurrence that are derived from these models are likely to under-estimate the true impact.

Of course, this is only one study, and it should also be noted that the most intense rainfall events are usually associated with tropical areas and with those in the path of hurricanes and in particular typhoons. There is a great deal more work to do on this topic, but the initial results provide real cause for concern.

Reference
Liu, S., Fu, C., Shiu, C., Chen, J., & Wu, F. (2009). Temperature dependence of global precipitation extremes Geophysical Research Letters, 36 (17) DOI: 10.1029/2009GL040218

Are satellite-based landslide hazard algorithms useful?

In some parts of the world, such as the Seattle area of the USA, wide area landslide warning systems are operated on the basis of rainfall thresholds. These are comparatively simple in essence - basically the combination of short term and long term rainfall that is needed to trigger landslides is determined, often using historical records of landslide events. A critical threshold is determined for the combination of these two rainfall amounts - so for example, it might require 100 mm of rainfall in hours after a dry spell, but 50 mm after a wet period. These threshold rainfall levels have been determined for many areas; indeed, there is even a website dedicated to the thresholds!

In 1997 NASA and JAXA launched a satellite known as TRMM (Tropical Rainfall Monitoring Mission), which uses a suite of sensors to measure rainfall in the tropical regions. Given that it orbits the Earth 16 times per day most tropical areas get pretty good coverage. A few years ago Bob Adler, Yang Hong and their colleagues started to work on the use of TRMM for landslide warnings using a modified version of rainfall thresholds. Most recently, this work has been developed by Dalia Bach Kirschbaum - and we have all watched the development of this project with great interest. The results have now been published in a paper (Kirschbaum et al. 2009) in the EGU journal Natural Hazards and Earth Systems Science - which is great because NHESS is an open access journal, meaning that you can download it for free from here.

Of course a rainfall threshold on its own doesn't tell you enough about the likelihood of a landslide. For example, it doesn't matter how hard it rains, if the area affected is in a flat, lowland plain then a landslide is not going to occur. To overcome this, the team generated a simple susceptibility index based upon weighted, normalised values of slope, soil type, soil texture, elevation, land cover and drainage density. The resulting susceptibility map is shown below, with landslides that occurred in 2003 and 2007 indicated on the map:


A simple rainfall threshold was then applied as shown below:

Thus, if an area is considered to have high landslide susceptibility and to lie above the threshold line shown above based upon an analysis using 3-hour data from TRMM, then a warning can be issued.

Kirschbaum et al. (2009) have analysed the results of their study using the landslide inventory datasets shown in the map above. Great care is needed in the interpretation of these datasets as they are derived primarily from media reports, which of course are heavily biased in many ways. Examination of the map above does show this - look for example at the number of landslide reports for the UK compared with New Zealand. The apparent number is much higher than in NZ, even though the latter is far more landslide prone. However, in New Zealand the population is small, the news media is lower profile, and landslides are an accepted part of life. However, so long as one is aware of these limitations then this is a reasonable starting point for analysing the effectiveness of the technique.

So, how did the technique do? Well, at a first look not so well:



In many cases the technique failed to forecast many of the landslides that actually occurred, whilst it also over-forecasted (i.e. forecasted landslides in areas in which there were none recorded) dramatically. However, one must bear in mind the limitations of the dataset. It is very possible that landslides occurred but were not recorded, so at least to a degree the real results are probably better than the paper indicates. Otherwise, the authors admit that the susceptibility tool is probably far too crude and the rainfall data to imprecise to get the level of precision that is required. However, against this one should note that the algorithm does very well (as indicated by the green pixels on the map above) in some of the key landslide-prone areas - e.g. along the Himalayan Arc, in Java, in SW India, the Philippines, the Rio de Janeiro area, parts of the Caribbean, and the mountains around the Chengdu basin. In places there is marked under-estimation - e.g. in Pakistan, Parts of Europe and N. America. In other places there was dramatic over-estimation, especially in the Amazon Basin, most of India, Central Africa and China.

All of this suggests that the algorithm is not ready for use as an operational landslide warning system. Against that though the approach does show some real promise. I suspect that an improved algorithm for susceptibility would help a great deal (maybe using the World bank Hotspots approach), perhaps together with a threshold that varies according to area (i.e. it is clear that the threshold rainfall for Taiwan is very different to that of the UK). Kirschbaum et al. (2009) have have produced a really interesting piece of work that represents a substantial step along the way. One can only hope that this is developed further and that, in due course, an improved version of TRMM is launched (preferably using a constellation of satellites to give better temporal and spatial coverage). That would of course be a far better use of resource than spending $4,500 million on the James Webb Space Telescope.

Reference
Kirschbaum, D. B., Adler, R., Hong, Y., and Lerner-Lam, A. 2009. Evaluation of a preliminary satellite-based landslide hazard algorithm using global landslide inventories. Natural Hazards and Earth System Science, 9, 673-686.

Future British seasonal precipitation extremes - implications for landslides

One of the great questions of the age is of course the ways in which climate change will affect the weather patterns that we are likely to see in the future. In the case of landslides the key issue is the ways in which precipitation patterns will alter, especially the most intensive rainfall events that are responsible for many of the most damaging landslides. One of the most significant steps forward over the last few years has been the ability of global climate models to handle these extreme events, meaning that at last we are starting to develop some capability.

This week an important paper has been published by Fowler and Ekstrom (2009), which seeks to look at the likely changes to very intense rainfall events in the UK. Helen Fowler, is based just up the road from me at Newcastle (the city with the chronically under-performing football team), and her co-author have used modelling ensembles to examine how UK precipitation regimes are likely to change in the time period 2070 to 2100 under the SRES A2 emissions scenario, which is currently effectively our best estimate as to how carbon dioxide emissions will change with time (Fig. 1).

Fig 1: SRES Emissions Scenarios. A2, as used in this study, is shown in Fig. (b). Source: http://www.grida.no/publications/other/ipcc_sr/?src=/Climate/ipcc/emission/014.htm

Ensemble modelling looks at the results of a series of different climate models to examine the range of outputs. Each model operates in a slightly different way, meaning that there will always be a range of results. Therefore, papers presenting ensemble model outcomes always present a range. One of the key issues of interest is whether there is some consistency between them. In this study. Of course the results of such modelling runs are highly complex - in this paper the authors have looked at the 1 day and 10 day precipitation events with a current return period of 25 years. The 1 day event can be thought of as the impact of an intense storm; the 10 day probably simulates a series of low pressure systems tracking across the country, as has happened several times in the last few of years. In landslide terms the 1 day storms might trigger the catastrophic debris flow and sallow failure events, whilst the 10 day events might trigger deeper seated and large slope failures.

First the model is run for the a control period (1961-1990) to check that they can realistically simulate observed conditions. They can. The models are then run to look at what would happen in the period between 2070 and 2100, and the results are then pooled using a fairly interesting approach. Well, the first thing to say is that the Global Climate Models (GCMs) do predict a much warmer climate - global mean temperatures are predicted to be 3.1 to 3.56 degrees warmer than at present. Interestingly though the occurrence of these intense rainfall events also greatly increases for three of the four seasons:
Winter: Increases in occurrence of extreme precipitation of 5 to 30%
Spring: Increases in occurrence of extreme precipitation of 10 to 25%
Summer: Very varied results, with some models suggesting decreases and other increases. More work is needed
Autumn (Fall): Increases in occurrence of extreme precipitation of 5 to 25%

A few of the models do predict larger (and smaller) increases - look at the paper for the full detail. Overall, the authors conclude that "Nevertheless, importantly for policy makers, the multi-model ensembles of change project increases in extreme precipitation for most UK regions in winter, spring and autumn. This change is physically consistent with warmer air in the future climate being able to hold more moisture. The use of multi-day extremes and return periods also showed that short-duration extreme precipitation is projected to increase more than longer-duration extreme precipitation, where the latter is associated with narrower uncertainty ranges."

The implications for landslides are stark. Increases on this level of the occurrence of extreme precipitation events will inevitably increase the occurrence of slope failures. Therefore, we should expect to see an increase in the occurrence of slope failures. Unfortunately, as landslides are triggered by just a small proportion of our existing rainstorm events, increases in this range are likely to have a disproportionate impact.

Of course the next thing to do will be to build the outputs of these models into slope stability models. This will be a fascinating exercise.

Reference:
H. J. Fowler, M. Ekström (2009). Multi-model ensemble estimates of climate change impacts on UK seasonal precipitation extremes International Journal of Climatology DOI: 10.1002/joc.1827

Global warming and landslide occurrence

One of the most vexed questions in landslide science at the moment is that of the potential link between climate change and mass movement occurrence. I have yet to meet a landslide researcher who does not believe in the reality of anthropogenic global warming, so we are all deeply interested in how our particular systems are likely to respond. Unfortunately this is not an easy question to answer for three reasons:

1. Landslides respond to changes in pore pressure (i.e. groundwater level). Groundwater level is controlled by precipitation input and by evapotranspiration outputs. So, to know how groundwater will respond requires quantification of both of these parameters. It might be expected that in a warmer world on average precipitation will increase (see below), but evapotranspiration will also increase. Understanding the balance between these parameters is at best a challenge.
2. Landslides are localised phenomena, usually sited in upland areas in which rainfall patterns are complex and variable. Unfortunately, global climate models work at much larger spatial scales (typically 1 or 1.5 degrees of latitude and longitude). This makes it difficult to scale the outputs to an individual landslide.
3. In many parts of the world, precipitation is controlled by large-scale weather systems, such as ENSO and the Asian SW monsoon. Climate models are struggling to model these systems adequately.

In the last month an interesting paper has been published that starts to take us in the right direction. The paper is this one:

Allan, R.P. and B.J. Soden, 2008. Atmospheric warming and the amplification of precipitation extremes. Science, 321 (5895), 1481-1483. You can download a copy from Brian Soden's website.

This paper is interesting because it looks at extreme precipitation in the context of climate change. In the last couple of years it has become clear that many of the most damaging landslide events tend to occur as a result of precipitation extremes - i.e. comparatively short duration, high intensity rainfall (the type associated with a particular storm or front) rather than long duration, lower intensity events. Understanding how extreme precipitation will change is thus very helpful.


Allen and Soden have started from the observation that the GCMs all forecast that extreme precipitation events will become more common as the climate warms. They have used a combination of measurements of daily precipitation over the tropical ocean using a NASA satellite instrument called SSM/I and the outputs from global climate models to look at the response of tropical precipitation events to natural changes in surface temperature and atmospheric moisture content. In the context of landslides, a very clear link was observed. In periods when temperatures were high the number of observed extreme rainfall events increased, and vice-versa. What is surprising though is that the response of the natural system seems to be more extreme than that of the climate models - i.e. the climate models are too conservative in terms of their forecasts of these extreme events.

The relevance of these findings for landslides should be quite clear. Increases in the occurrence of extreme precipitation intensities might well be expected to increase the occurrence of landslides. It should be noted though that there is some way to go to really establish this link. For example, this paper is essentially based on a dataset collected over the ocean. There is a need to see whether the same applies on land. However, it is good to see papers being produced that start to answer the questions that we are asking.