ABSTRACT
Barbier (2007) has advocated the use of EDF to measure the storm protection value of coastal wetlands. This methodology is based on the concept of compensating surplus that a representative household requires to maintain its utility level due to the expected increase in the incidence of damaging storm events. The coastal wetlands are assumed to have a direct effect on reducing the production of damaging natural disasters in terms of their ability to inflict damages locally. Barbier assumed that the presence of wetlands in some areas will reduce the occurrences of damaging storm events (as they provide protection) and thus would also reduce the amount of compensating variation needed to be paid to the household to maintain their utility level. Hence, the change in consumer surplus due to change in wetland area will measure the storm protection value of the wetland. However, storm occurrences and their intensities are dependent on climatic conditions, and the damage occurrences due to the storm at a place depend on both the storm features and the physical features of the province. The wetland area of a province may reduce the occurrences of actual damages compared to the potential damages expected from a cyclone, but the presence of wetlands only will have little influence on the ranking of the storm in terms of the damages they caused or the number of damaging storms hitting the province. One has to take into account the storm characteristics as well as the other features of the location. As explained below, storm damages are dependant on multiple factors, and wetland areas are only one among many others. Moreover, such analyses are needed at a micro level, rather than at the level of a province, as the effects of the wetlands are more localized and location-specific. The use of the avoided damages approach for storm protection started with Farber’s (1987) pioneering work in which he used a scientific model of wind velocity and valued the protection value of wetlands from only the perspective of the wind damage from hurricanes. Farber talked about the homogeneity of the population, and this may justify the exclusion of socio-economic factors from damage analysis. But the limitation of his work was that he assumed that wetlands reduce the wind damage, whereas wetlands – unless they have high vegetation – can provide little protection from wind. Of course, wetlands do provide distance as a buffer between the coast and human settlements, which may result in the wind speed being lower by the time the hurricane strikes the settlements. However, if the wetlands are entirely water bodies, then the weakening of the storm becomes less prominent due to moisture supply than it would have been with a more rough surface. Wetlands provide more protection from storm surges and Farber did not account for storm surges in his model. Costanza et al. (1989, 1997, 2008) talked about the storm protection value of wetlands, but the values used in their earlier papers are taken from elsewhere. The latest study estimates the storm protection value of the coastal wetlands of the United States by using the avoided damages approach. The authors estimated a regression model for 34 major US hurricanes since 1980, with the natural log of damage per unit of gross domestic product in the hurricane swath as the dependent variable, and the natural logs of wind speed and wetland area in the
swath as the independent variables. Storm protection value is estimated from the marginal effects. Along with wind speed and wetland area, storm damages depend on multiple other factors and thus the results of the study could be suffering from omitted variable biases. Among the analytically best studies on the protection of mangroves from cyclone damages has been the well-conceptualized work by Badola and Hussain (2005). Conducting primary surveys for damages in the aftermath of the super cyclone of October 1999, they showed the damages per household to be less in a village sheltered by mangroves compared to the damages per household in a village having a dike nearby but no mangroves, and a village without either mangroves or dikes. Though the authors are given credit for selecting villages as similar as possible – except for the presence or absence of mangroves and dikes – in order to neutralize the impacts of socio-economic factors on storm damages, these impacts are still visible from the summary statistics. However, the attribution of the entire reduced damages by the authors to mangrove presence appears to be a biased overestimate:
1 The study villages are not socio-economically homogeneous – a glance at the general characteristics of the villages shows their economic heterogeneity. Hence, a statistical analysis looking at the effect of all the factors simultaneously on damage occurrences will definitely give different results.
