As is now the established format the Annual Flood Report is made up of three major sections:
The annual theme, which for 2103 is the regional incidence of typhoons and tropical storms and their role in the flood hydrology of the Lower Mekong Basin.
A review of the flood season over the year, and
A summary overview of the four National Flood Reports
In an average year 4 to 6 typhoons or severe tropical storms make landfall in Viet Nam, a number of which will track across the Lower Mekong Basin and cause significant to extreme flooding. There are many years when far more systems make landfall, for example 1964 (18), 1973 (12), 1978 (12), 1989 (10) and 1996 (10). The incursion of these storm systems into the Mekong region has historically been associated with most of the largest flood peak discharges on the mainstream. This said, it has to be acknowledged that the annual Mekong flood is in its greater part a response to the SW Monsoon and is a multivariate event defined by not only the seasonal maximum discharge but also by the volume of floodwater and the duration of flows above critical thresholds. The events of 2000 illustrated quite clearly that extreme floods cannot be defined exclusively in terms of the annual maximum discharge. On that occasion the flood peak was no more than average but the volume of floodwater over a prolonged flood season was critical and had a devastating impact across the Cambodian floodplain and within the Mekong Delta.
The role of typhoons and tropical depressions in the flood hydrology of the Mekong is well established. However, there has never been a detailed systematic study nor an inventory of their annual incidence and impacts. The challenge lies with the fact that in meteorological terms the severity of tropical low pressure systems is indicated according to wind speed, while in a hydrological context the focus of interest lies with the consequent storm rainfall and resulting flood runoff. Prior to the introduction of the Annual Flood Reports by the FMMP in 2006/7, supported by the four National Reports, assembling the relevant historical data is a considerable task since ideally storm rainfall maps are required for each of the key events. The further back in time one goes the more difficult it is to put together sufficient rainfall data to accurately depict the geography and intensity of the storm rainfall. This arises because the observation network becomes increasingly sparse.
The exercise undertaken here has proved to be useful and informative, with the linkage between tropical low pressure systems and regional flooding examined back to the early 1950’s. In a sense though, it should be seen as “exploratory” and setting the framework for a more detailed research assessment. The importance of undertaking the latter lies within the context of potential climate change impacts upon regional floods and flooding. Due to warming sea temperatures the incidence and severity of tropical low pressure systems is forecast to intensify with secondary consequences with regard to the frequency of intense storms and flooding. As is made evident in Section 2, there is no evidence to suggest that the regional annual count of storm systems has increased in recent years, although it may be the case that their scale and severity has intensified, or at least the occurrence of super typhoons such as HAIYAN in 2013, has become more frequent. This though could be difficult to establish to any degree of statistical satisfaction. Super typhoons and extreme tropical storms are hardly a contemporary development. Chinese historical annals and Vietnamese records chronicle a long history of tropical storm impacts. Amongst these is the Haiphong super typhoon of 1881 which killed a reported 300 000 people and is regarded as the third most deadly tropical storm in recorded world history.
If the intensity of tropical storms has increased in recent years, the evidence would lie in part with systematically higher short duration (one to three days) storm rainfall over modern decades. As shown in Section 2 there is no regional evidence to suggest that this is the case.
The WMO data examined in Section 2.1 quite clearly indicates an increase in Asian storm and flood related disasters in each decade since 1970. The question that arises in this context is, however, is whether the increase in the rate of disasters and the associated fatalities and economic damage is a consequence of more frequent extreme events brought on by climate change or whether it is simply a case that more people and infrastructure are exposed. The latter argument is compelling. In Asia over the last 50 and more years population increase has been historically unprecedented and the subsequent pressure on land and agricultural resources has forced the settlement of exposed and vulnerable sub regions, particularly in river deltas and to a lesser extent in marginal upland areas. In effect the susceptibility of the regional populations to meteorological and hydrological disasters has grown relentlessly.
An issue with the WMO study, which is acknowledged, is that the results are determined in large part by just a few decisive events. These include the Bangladesh cyclones of 1970 and 1991 which killed almost 450 000 people between them, cyclone NARGIS in Myanmar in 2008 which caused over 136 000 deaths, floods in Thailand in 2011 which caused US$ 41 billion in damage and a tropical cyclone in Japan in 1991 which caused US 17 billion in damage and was the costliest on record.
Despite the influence of these prominent events on the statistics, the WMO findings describe a disturbing trend. Whether this is climate driven in some measure or far more the consequence of socio economic pressure on resources is arguable.
The climate change debate, in terms of systematic quantitative evidence, revolves around increasing maximum air temperatures, sea temperature, sea levels and polar ice coverage. These indicators can be monitored on a year by year basis as a sequence of continuous random variables. Historic trends are widely accepted as established. Much more challenging is the statistical evaluation of discrete variables, that is events such tropical storms and floods, which occur in discrete or disconnected points in time. Statistically, they are defined as a “point process”. The components of interest are the annual “count” and their intensity above a predescribed threshold level. The statistical analysis of such data is complex compared to that of continuous time series. Within the field of the earth sciences very little research has been carried out in order to establish whether significant changes to the incidence and severity of tropical storms, for example, is evident.
Consequently, the case for potential future change is based upon the physical effect of increasing sea surface temperatures and climate modeling. The possible impacts upon the storm / flood linkage is then assessed on the basis of “what if” scenario analysis, again using numerical models. A major issue though, as has already been indicated, is the increasing vulnerability of riparian societies to flood induced fatalities, loss and damage, especially those living on the margins of economic development. What clearly emerges is that yesterday’s norms will not be the same as tomorrow’s.
However, historical, geo-referenced information about deaths and damages can be used to estimate risks before the next disaster occurs. It can support practical measures to reduce potential impacts, such as investing in early warning systems, retrofitting critical infrastructure or enforcing new building codes. Information about past impacts can also be used to assess the resilience of a society.