Nature Conservancy, 2 Collaborators Issue Issues Report Entitled 'Reducing Caribbean Risk – Opportunities for Cost-Effective Mangrove Restoration and Insurance'(Part 2 of 4)

WASHINGTON, April 17 (TNSrep) -- Three organizations issued a 34-page report in April 2023 entitled "Reducing Caribbean Risk: Opportunities for Cost-Effective Mangrove Restoration and Insurance." The three organizations were the Nature Conservancy, University of California Santa Cruz and AXA.

Here are excerpts:

(Continued from Part 1 of 4)

* * *

Benefit-Cost Analysis of Mangroves in the Caribbean

Government agencies and the world's biggest (re-)insurers are considering how their funds could be invested in habitat restoration to reduce future risk and build resilience. For example, the US Federal Emergency Management Agency (FEMA) has indicated that they can fund restoration with storm recovery funding if it can be shown in their benefit-cost analysis that such projects achieve benefit-cost ratios (B:C) greater than 1. Likewise, the insurance industry and ecologists are developing tools to estimate where resilience insurance could be used to pay for restoration up front based on reduced risk and premiums later (Reguero et al., 2020). To realize these opportunities, we must provide critical information on mangrove benefits and restoration costs to build the B:C maps that can guide cost effective mangrove conservation and restoration.

Methods

We have developed benefit-cost maps based on (i) recently published work that measures the benefits of mangroves for flood risk reduction globally (Menendez et al., 2020), (ii) new data on the costs of mangrove restoration, and (iii) the assessment of mangroves as a 30-year coastal asset. The core assessment considers coastal flood risk and the value of mangroves for reducing this risk.

To measure and value the coastal protection benefits provided by mangroves, we follow the Expected Damage Function approach (Figure 3) commonly used in engineering and insurance sectors and recommended for the assessment of coastal protection services from habitats (World Bank, 2016). The flood protection benefits provided by mangroves are assessed as the flood damages avoided to people and property by keeping mangroves in place. We couple offshore storm models with coastal process and flood models to measure the flooding that occurs with and without mangroves under different storm conditions. These flood extents are used to estimate the averted flood damages to people and property and hence the expected benefits of mangroves in social (people protected) and economic terms (value of property protected). These methods have been applied in a number of prior projects to assess the value of coral reefs for coastal protection globally (Beck et al., 2018a), and to assess the value of mangroves for coastal protection in the Philippines, Jamaica, and globally (Losada et al., 2017; Menendez et al., 2018; Menendez et al., 2020; Beck et al., 2019a).

Our estimates are based on a set of global process-based models, applied to the Caribbean region, that identify the annual expected benefits of mangroves for flood risk reduction (Menendez et al., 2020). We use a set of hydrodynamic and economic models to identify the area and depth of flooding (i) for mangrove coastlines, (ii) in model runs with and without mangrove, and (iii) for four storm return periods, 1 in 10, 25, 50, 100-year driven by local storm data. We first develop and validate these models in the Philippines, a country with over 36,000 km of heavily populated coastlines, at high risk from cyclones, and more than 200,000 hectares of coastal mangroves (Menendez et al., 2018). We use these models to develop a dataset of several thousand simulations to statistically describe the physical relationships between (i) tropical cyclones and offshore wave climate; and (ii) offshore wave climate, mangrove parameters, and extreme sea level in coast. This dataset is then used to build regression models and interpolation tables to estimate how mangroves modify extreme water levels at the shoreline. Finally, for every kilometer of mangrove shoreline globally, we overlay the resulting coastal lood maps on economic asset information downscaled to 30 x 30 meters and assessed by flood depth to identify flood damages (risk) and avoided damages (mangrove benefits).

The values that we have provided for both mangroves and coral reefs (Beck et al., 2018a; Menendez et al., 2020) are the first global estimates of flood risk reduction benefits provided from process-based models for any coastal or marine ecosystem. This work represents a state of the art in global flood risk and benefits assessment and has been shown to provide better estimates than replacement cost approaches (Barbier et al., 2015; World Bank, 2016). We have chosen this approach over others because it is (i) quantitative in contrast to other approaches, which use indicator (expert) scores to assess shoreline vulnerability (e.g., Silver et al., 2019), (ii) it uses the methods and tools of risk agencies, insurers, and engineers (Narayan et al., 2016; Narayan et al., 2017; Reguero et al., 2018), (iii) it is consistent with approaches for national accounting (World Bank, 2016), and (iv) it accurately captures impacts of extreme cyclone events that are typically underestimated in global studies.

We measure the flood protection benefits of mangroves all over the world for coastal flooding from extreme water levels at the shoreline. Following this approach, the role of mangroves in coastal protection is examined by measuring the economic impacts of coastal flooding on people and property under two scenarios: with and without mangroves. The "without mangroves" scenario assumes complete loss of the habitat and the consequent erosion of the intertidal area into a smooth sandy surface.

Our global study covers 700,000 km of mangrove-inhabited coastline that includes more than 141,000 square km of mangroves. The models require a huge amount of data and high computational effort, so a four-level subdivision of the world is made to organize data and modeling efforts (see Figure 4). The first level is the division into five macro-regions, corresponding to the five ocean basins of tropical cyclone generation (Knapp et al., 2010). The second level divides the 700,000 km of coastline into 68 sub-regions considering coastline transects of similar coastal typology (e.g., islands and continental coasts) and similar ecosystem characteristics. The third level of disaggregation involves local scaling, taking units with 20 km of coastline and extending up to 30 km inland and 10 km seaward, with the aim of providing local results. The fourth, and final, subdivision is the coastline profiling of every kilometer.

* * *

Figure 3: Key Steps and Data for Estimating the Flood Protection Benefits Provided by Mangroves.

Figure 4: The Geographic Subdivisions for Hydrodynamic Models.

* * *

First, we define cross-shore (i.e., seaward to landward) coastal profiles every 1 km for all mangrove coastlines globally and group them into 20 km study units. The 700,000 global coastal profiles are grouped and classified using a library of 750 pre-existing cross-shore, 1-D profiles that were developed based on data from the Philippines. Each profile contains the following information: (i) profile slope (i.e., from mean water depth along the profile at multiple distance intervals from offshore to shoreline), and (ii) total height and width of mangroves.

Then, we follow a multi-step framework:

1. Estimate offshore dynamics. Produced from tropical cyclones and regular climate conditions. This measures the flood protection service of mangroves all over the world for two climatic conditions: (i) Cyclonic, i.e., the conditions of high-intensity extreme waves and storm surge induced by tropical cyclones; and (ii) Non-cyclonic, i.e., the "regular" waves generated by low-intensity local storms. The data sets on tropical cyclones and waves are global and provide locally specific information from more than 7,000 historical cyclones (Knapp et al., 2010) and 32 years of data on waves and sea level.

2 Estimate nearshore dynamics produced by non-cyclonic and cyclonic conditions. Once we resolve offshore dynamics, we obtain waves and storm surge on the seaward side of each cross-shore profile. Waves interact with the sea floor and other obstacles (e.g., islands) as they approach the coast and modify height and direction through shoaling, refraction, diffraction, and breaking processes. Regular climate is propagated following a hybrid downscaling. The 32-year long series, from 1979 to 2010, includes 280,000 sea states (one sea state is a 1-hour record of wave height, peak period, and total water level). Considering the 700,000 coastal profiles and the 280,000 sea states results together is an unmanageable number of cases. Therefore, we reduce the number of sea-state propagations by considering only the 3,787 non-repeated combinations of wave height, peak period and total water level (SS+AT+MSL) and, then, applying the Maximum Dissimilarity Algorithm (MDA) to obtain 120 sea states to be propagated to shore following with Snell's law and the shoaling equation. Tropical cyclone nearshore hydrodynamics are estimated using a previously derived regression model (see Menendez et al., 2020). We apply regression models in each profile, and we obtain the same parameters as for regular climate.

3. Estimate the effects of mangroves on flood reduction. This consists of propagating ocean hydrodynamics over mangrove forests which dissipate wave and surge energy, and, consequently, reduce flood height. Flood height is a function of mean sea level, astronomical tide and wave runup. Mangrove dissipation takes place by means of breaking and friction processes. Given the large scale of this analysis, we follow a simplified approach for vegetation modeling. We use the model developed in the Philippines to infer the flood height (output) given mangrove forest width, significant wave height, peak period, and total water level (inputs). Then, we apply the statistical reconstruction technique RBF (Radial Basis Functions) to calculate the complete historical flood height time series at each profile. Next, we carry out an extreme value analysis. First, we select maximum values on a variable threshold (minimum, 1-in-5-year event). We adjust these selected values to a Generalized Pareto-Poisson distribution, and we obtain the flood height vs return period curves for both scenarios: with and without mangroves.

4. Estimate the land flooded (impact) due to extreme water levels along the shoreline by intersecting the flood height with topography. We use a global GIS model to calculate the extent and height of inland flooding due to the flood height at the shoreline with and without mangroves, from tropical cyclones and regular climate.

5. Calculate the (averted) flood damages to people and property. We use global datasets and GIS models to estimate the damages to property (economic) and people (social) from flooding due to tropical cyclones and regular climate, with and without mangroves. We determine flood damage using depth-damage curves, which identify the flood damage that would occur at specific water depths. Two sources of information have been used to obtain these damage curves: the EU Joint Research Centre (JRC) (Huizinga et al., 2017) and US Hazus (Scawthorn et al., 2006). Hazus is based only on US collected data but frequently extrapolated for use in other geographies. Whenever possible, we used the new curves from JRC, which used an extensive literature survey to develop damage curves for each continent with some additional differentiation between countries to establish maximum damage values.

Data and Model Assumptions

Population Data

Global exposure data for people was obtained from the Socioeconomic Data and Applications Center (SEDAC) fourth version of Gridded Population of the World at a 1 km spatial resolution (http://sedac.ciesin.columbia.edu/data/collection/gpw-v4). SEDAC is freely available, and includes a map viewer to see the global distribution of different socio-economic assets (http://sedac.ciesin.columbia.edu/mapping/viewer/).

Residential and Industrial Property Data

This study uses data from GAR15 (Desai et al., 2015) on the economic value of the residential and industrial building stock, which is based on 2010 economic data from the World Bank (World Bank, 2011; De Bono and Chatenoux, 2015). Throughout this report, we use stock and property interchangeably to mean the physical buildings. The GAR15 provides a global exposure database with a standard 5 km spatial resolution and a 1 km detailed spatial resolution on coastal areas, estimating the economic value of the exposed assets, as well as their physical characteristics in urban and rural agglomerations. The variables included in the database are number of residents, and economic value of residential, commercial and industrial buildings (De Bono and Chatenoux, 2015). The GAR15 database follows a top-down approach using the geographic distribution of population and gross domestic product (GDP) as proxies to distribute the rest of the socio-economic variables (population, income, education, health, and building types) where statistical information including socio-economic, building type, and capital stock at a national level are transposed onto the grids of 5x5 km or 1x1 km using geographic distribution of population data and GDP as proxies (UNISDR, 2015). We downscaled residential and industrial stock data from the GAR15 using the population raster (from WorldPop, 100m resolution).

Gross Domestic Product

World Development Indicators from the World Bank (https://datacatalog.worldbank.org/ dataset/world-development-indicators) were used to obtain GDP data for each country involved in this study (World Bank, 2017). GDP information is available from 1960 to 2016. Additionally, World Bank databases were used to validate other data-sources: population from SEDAC and residential and industrial stock from GAR15.

Damage Functions

Global flood depth-damage functions are needed to evaluate damages for different flood levels. A new report from the EU Joint Research Centre (JRC) collected data from Africa, Asia, Oceania, North America, South America and Central America and proposed damage functions for residential and industrial stock, commerce, transport, infrastructure and agriculture (henceforth, JRC damage) at each location (Huizinga et al., 2017). These damage functions are a new alternative to damage curves from Hazus databases (Scawthorn et al., 2006). JRC damage functions are intended to address flooding effects on property globally, developing a consistent database of depth-damage curves.

Limitations and Adjustments

Our efforts represent state of the art process-based assessments of flood risk and mangrove benefits globally. For most countries with mangroves, these represent the best data and models for mangrove benefits, and for many countries the best national level estimate of flood risk. For this study, we have developed a dataset of several thousand simulations to describe how mangroves modify extreme water levels at the shoreline, for every kilometer of mangrove coastline in the world. This approach is computationally highly efficient and allows us to estimate coastal flood risk for new scenarios of mangrove presence and extent. However, for local scale analyses, it is sometimes possible to obtain higher resolution data for example for bathymetry, topography, and assets.

Based on prior work and our own sensitivity analyses, the greatest sources of uncertainty in coastal flood risk assessments are estimates of topography (Menendez et al., 2019). Given that flooding and damage from tropical storms are among the greatest risks to people and property, better elevation and depth data is urgently needed. Fortunately, in the past decade, there has been a substantial increase in the availability of high-resolution coastal elevation data through the widespread use of LIDAR. Nearshore bathymetry, however, remains a major gap, though there are advances in remote sensing that could help.

Our coastal flooding analyses have several significant, combined improvements over other recent global flooding analyses including downscaling to 30 m resolution; consideration of hydraulic connectivity in the flooding of land; the use of 30 years of wave, surge, tide, and sea level data; reconstruction of the flooding height time series and associated flood return periods. Our flood risk models also include ecosystems for the first time, which represent critical advances in the assessment of flood risk. Major remaining constraints for global coastal flooding models include the consideration of flooding as a one-dimensional process and the difficulty in adequately representing flooding on smaller islands.

Our preliminary review of the results from the global analysis identified that a handful of countries had very high values of benefits per hectare (up to millions per hectare). To be conservative, we assumed that these values were too high and represented outliers. Two measures were taken to address these outliers. First, countries with less than 100 hectares of mangroves were excluded from the analyses as there were too few mangroves in these countries to reliably estimate benefits from a global model. This excluded a total of 15 countries including Bahrain, Singapore, and Benin as well as eight Caribbean Small Island Developing States.

Once we excluded the countries above, there were 7 more countries including the US, China, and Vietnam with benefits per hectare that were considered exceptionally high. For these countries, we capped the estimated benefits per country from our global model at $50,000 per hectare. This value was based on the maximum benefits per hectare from high resolution flood risk and mangrove benefits that we calculated using the best insurance industry risk models and capital exposure datasets (see for example Narayan et al., 2017). Specifically, in work done with Risk Management Solutions (RMS), a leading insurance risk modeling firm, we found the benefits of Florida mangroves during Hurricane Irma to be as high as $47,000 per hectare at the county level (Narayan et al., 2019).

It is certainly possible to design specific mangrove restoration projects to deliver very high flood protection benefits (i.e., much greater than $50,000 per hectare). Examples of high value restorations could be a few hectares of mangroves placed in front of airports, port facilities, bridges or high value homes and condos. However, for the purposes of these analyses across 20 km stretches of coastline, a cap of $50,000 per hectare represents a conservative estimate of benefits. The purpose of this study was to identify areas with benefit to cost ratios greater than one (B:C>1); all coastal study units with benefits at $50,000 per hectare had very high B:C ratios.

* * *

Assessing Costs of Mangrove Restoration

We gathered published data on the costs of mangrove restoration across the wider Caribbean region. We also identify factors that are particularly important in determining the costs of mangrove restoration projects. In total, we assess data from 137 mangrove restoration projects world-wide, including 72 projects implemented in the Caribbean (Narayan et al., 2019). These data are obtained through a systematic literature review of the reported costs of mangrove restoration projects in Jamaica and the Caribbean region, and the costs of coastal protection structures in Jamaica, using Google, Google Scholar, Web of Science, and Scopus search engines. We extend and build on the data provided by the comprehensive review conducted by Bayraktarov et al. (2016) and Narayan et al. (2016). All cost data were combined with information on project areas to obtain a cost per hectare. In addition to the literature review, we reached out to relevant government and other institutions in Florida, Jamaica, and the wider Caribbean for information on any site-specific factors that would influence these costs including project areas, restoration techniques, and individual project costs.

Restoration costs across the wider Caribbean are generally comparable and vary from around $23,000 per hectare in countries like Guyana to around $14,000 in Grenada (Figure 5). The costliest location in the Caribbean region for mangrove restoration is Florida, with median costs of $45,000 per hectare and extreme variability. We also report on restoration costs in other regions, which primarily represents data gathered from SE Asia where restoration can be cheaper than in the Caribbean (see Figure 5).

Costs per hectare are typically lower for larger restoration projects. In general, the factors influencing the costs of mangrove restoration projects are four-fold: (i) the costs of land and permits; (ii) the costs of obtaining and transporting the material; (iii) the costs of designing and constructing the project, and; (iv) the costs of monitoring and maintaining the project post-construction (Narayan et al., 2019). Since mangrove restoration happens in the inter-tidal zone, the availability and price of land and the necessary permits are an important factor influencing costs. Another factor that influences costs is the restoration technique. Restoration by planting mangrove saplings manually can be cheap if these projects make use of local, voluntary labor. Projects involving hydrological restoration can be more expensive due to the need for specialized equipment, labor and the purchase and transportation of sediment. Maintenance and monitoring are other important cost components, though often not reported in restoration projects. We find that specific maintenance actions, such as fencing restoration sites to reduce disturbance can significantly add to overall project costs (Narayan et al., 2019).

* * *

Figure 5 Costs of Mangrove Restoration and Construction of Artificial Coastal Structures in the Caribbean and Other Regions. For mangroves, costs shown are per hectare. Number of studies, N, indicated in parentheses. All numbers are median costs, unless N=1. All costs are in 2019 USD$ and rounded off to the nearest 1,000.

Restoration costs vary by region, given the variation in the costs of project components. Restoration in the US is considerably more expensive than across the wider Caribbean, partly due to the higher costs of land, equipment, labor, and permits.

Thus, in our benefit to cost ratio analyses, we used two different average costs of restoration. For projects in the US, we assumed an average cost of restoration per hectare of $45,000. For projects across the rest of the Caribbean, we used an average cost of $23,000 per hectare.

Benefit-Cost Ratios

Our B:C analyses combine information from (i) annual expected flood risk reduction benefits ($) provided by mangroves in each 20 km coastal unit; (ii) total hectares of mangrove in each unit; and (iii) average costs of restoration per hectare.

Using data on mangrove benefits and restoration costs, we calculated benefit-cost ratios for each coastal unit. In our analyses, we assume that restoration means the return or recovery of mangrove habitat into areas where they once occurred. In addition, we calculated benefits per mangrove hectare for each coastal segment. We mapped the data in ArcGIS to visualize spatial differences in B:C ratios and benefit value per hectare.

To estimate B:C ratios, we assume that future restoration benefits per hectare will be similar to current flood risk reduction benefits per hectare within each coastal unit as measured in Menendez et al. (2020). Mangroves in areas with significant storms and high economic exposure offer more benefits per hectare than areas with fewer storms and less economic exposure.

We assume that mangrove restoration projects represent a 30-year coastal infrastructure asset (i.e., a 30-year project lifetime). We apply two different discount rates across this project lifetime: 4% and 7%. Four percent is consistent with values for project assessments with the World Bank. Seven percent is the required discount rate for projects assessed/supported by FEMA in the US.

Results

The results identify that there are cost effective opportunities for mangrove restoration across the Caribbean. There are 20 territories and countries in the region that have sections of coastline (i.e., ~20 km coastal study units) with cost effective opportunities for mangrove restoration (Figure 6). In total, we identified more than 180 coastal units (i.e., > 3,000 km of coastline) that have cost effective opportunities for mangrove restoration at a 4% discount rate. Cuba (37), Bahamas (23), and Florida (23) have the most study units with cost effective opportunities for mangrove restoration. The additional territories and countries rounding out the top 10 with the highest numbers of cost effective coastal study units (i.e., lengths of coastline with cost effective opportunities for mangrove restoration) were, in order, Mexico, Venezuela, Puerto Rico, Belize, the Dominican Republic, Jamaica and Guyana.

Results are robust to changes in discount rates. For mangroves, there are only 15 (8%) coastal study units that drop below the cost-effective threshold at a 7% discount rate (Figure 8) as compared to 4% rate.

Restoration project benefits per hectare varied widely across the Caribbean (Figures 7 and 9). These benefits per hectare results (Figures 7 and 9) identify where to find the expected break-even costs for mangrove restoration across the Caribbean; e.g., all areas in orange could achieve 1:1 B:C ratios for projects if restoration costs were $10,000 per hectare. The highest potential average restoration project benefits per hectare are in Antigua and Barbuda, the Dominican Republic, Saint Kitts and Nevis, and the British Virgin Islands.

* * *

Figure 6: Benefit to Cost Ratios for Mangrove Restoration across the Caribbean at 4% Discount Rate. Values are the net present value of restored mangroves as an infrastructure asset assuming a 30-year project with a 4% discount rate. Benefit values are based on Menendez et al. 2020. We assume a restoration cost of $45,000/ha for projects in the US and $23,000/ha across the rest of the Caribbean. The B:C ratios are summarized in coastal study units or blocks which cover approximately 20 km of coastline (see methods).

Figure 7: Benefits of Mangrove Restoration per Hectare at 4% Discount Rate. The values are summarized in coastal study units which cover approximately 20 km of coastline. We use the spatially explicit, annual expected flood risk reduction benefits for each coastal unit from Menendez et al. 2020 and divide by the total hectares of mangrove in each unit. We then calculate the benefits of the asset over a 30-year time period at a 4% discount rate.

Figure 8: Benefit to Cost Ratios for Mangrove Restoration across the Caribbean at 7% Discount Rate. Values are the net present value of restored mangroves as an infrastructure asset assuming a 30-year project with a 7% discount rate. Benefit values are based on Menendez et al. 2020. We assume a restoration cost of $45,000/ha for projects in the US and $23,000/ha across the rest of the Caribbean. The B:C ratios are summarized in coastal study units or blocks which cover approximately 20 km of coastline (see methods).

Figure 9: Benefits of Mangrove Restoration per Hectare at 7% Discount Rate. The values are summarized in coastal study units which cover approximately 20 km of coastline. We use the spatially explicit, annual expected flood risk reduction benefits for each coastal unit from Menendez et al. 2020 and divide by the total hectares of mangrove in each unit. We then calculate the benefits of the asset over a 30-year time period at a 7% discount rate.

Figure 10: Benefit to Cost Ratios for Private Property Benefits of Mangrove Restoration. This map considers only the flood reduction benefits in averted damages to private property across the Caribbean. Values are the net present value of restored mangroves as an infrastructure asset assuming a 30-year project with a 4% discount rate.

Figure 11: Benefit to Cost Ratios for Public Property Benefits of Mangrove Restoration. This map considers only the flood reduction benefits in averted damages to public property across the Caribbean. Values are the net present value of restored mangroves as an infrastructure asset assuming a 30-year project with a 4% discount rate.

* * *

(Continues with Part 3 of 4)

The report is posted at: https://www.nature.org/content/dam/tnc/nature/en/documents/TNC_MangroveInsurance_Final.pdf

TARGETED NEWS SERVICE (founded 2004) features non-partisan 'edited journalism' news briefs and information for news organizations, public policy groups and individuals; as well as 'gathered' public policy information, including news releases, reports, speeches. For more information contact MYRON STRUCK, editor, editor@targetednews.com, Springfield, Virginia; 703/304-1897; https://targetednews.com