Regardless of the politics surrounding climate change, the sea level is rising. Based on

Transcription

1 Analysis of Sea Level Rise Impacts in Bath, Bowdoinham, and Topsham, Maine Hannah Glover with Cam Adams, Liza LePage, and Dan Lesser ES204: Johnson 14 December 2012 INTRODUCTION Regardless of the politics surrounding climate change, the sea level is rising. Based on current predictions from the IPCC, sea level will rise.5 to 1.9m ( ft) in the next 100 years (Camill et al, 2012). In addition to rising sea levels, the number and severity of large storms has increased, bring damaging storm surge. The Patriot s Day Storm, Hurricane Katrina, and Hurricane Sandy have woken planners to the growing danger of inundation and flooding in coastal communities. Large cities have already begun to plan both for inundation and more frequent storm surges (Slovinski, 2012). The ability and desire to plan for these situations has been smaller in suburban, coastal communities like Mid-coast Maine. The semi-coastal towns of Topsham, Bowdoinham, and Bath in Sagadahoc County expressed interest in assessing the impacts of inundation on a small scale. The purpose of this EOS 204 community project was to provide these towns with the data and maps to push for further analysis. We examined the effects of sea level rise and storm surge on infrastructure in each town. Inundation levels of 1, 2, 3, 4 and 6ft were calculated on top of the Highest Annual Tide. These levels represent any combination of sea level rise and storm surge; for example a three foot increase could be 3ft of sea level rise or a storm surge of 2ft on top of a 1ft sea level rise. Maine is currently planning for 2ft of sea level rise so this study focused on 2ft sea level rise and a 2ft storm surge on top of that (4ft). The towns were primarily interested in the impact on infrastructure and habitat; we focused on roads, buildings, parcels, shoreland zoning, and waterfowl habitat; I analyzed

2 affected roads and building. The town planners will use the analysis to identify red flags and low hanging fruit (Slovinsky, 2012). Red flags are infrastructures such as water treatment plants and hospitals that would have disasterous consequences if inundated. Low hanging fruit are cheap, noninvasive changes that strengthen the shoreline against flooding. For example, by expanding the shoreland zoning, planners can require that more homes have flood insurance and are built to withstand flooding. All three of these towns have access to and some familiarity with ArcGIS. However they don t have dedicated staff working with sea level rise data. This project will be used for general planning purposes and to garner support for more analysis. The maps will be shown in town hall meetings to residents to increase awareness of the risks associated with sea level rise. The town planners will want access to all of the data we produce as well as the maps so that this project can support further work. DATA SOURCES The sea level rise layers were created from LiDAR DEMs. The majority of the towns were covered by FEMA LiDAR, created in This data was very clean and had accurate metadata. The portion of Topsham next to the Androscoggin was not covered by this LiDAR so NE Sagadahoc County LiDAR was used. The vertical projection on this data was labeled inaccurately and the metadata was unhelpful. This LiDAR had to be extensively manipulated before it could be combined with the FEMA LiDAR. Eventually these raster files were converted to polygons of individual sea level rises. The current sea level height was calculated from tidal information from buoys in Merrymeeting Bay ( The roads (E911), wetland habitat,

3 waterfowel habitat, and town boundaries were taken from the Maine Office of GIS Database. The parcel and tax assessor data, buildings points and outlines, and shoreland zoning were vector files from the individual towns. The rest of the data was created during the analysis process. ANALYSIS The first step in analyzing sea level rise was defining current sea level. The Highest Annual Tide was selected because it is a baseline model for the worst case scenario. This analysis shows the inundation that will happen once a year at the highest tide and what would happen if a storm hit during this tide. These levels are not the same for each town because the tide flows unevently into Merrymeeting Bay. NOAA and the Maine DEP run tidal gauges through the area and the HAT value was taken from the station closest to each town. The HAT and each sea level rise was calculated in meters in Table 1. The values were converted to meters for the UTM projection on the LiDAR. The HAT was not interpolated between these tidal buoys. HAT (ft) HAT (m) 1ft rise (in m) 2ft rise (in m) 2ft w/ 2ft storm surge (in m) 3ft (in m) 6ft (in m) Bath Bowdoinham Topsham Table 1. Highest Annual Tide and Sea Level Rise values for Bath, Bowdoinham, and Topsham. Combining the NE LiDAR to the FEMA LiDAR was a challenge because the NE LiDAR did not have the same vertical projection as the FEMA and furthermore it was mislabeled in the metadata. After experimenting with a variety of troubleshooting techniques, Eileen Johnson helped us reclassify the NE LiDAR so that the gridcodes were

4 correct and then the projection could be fixed. Once this obstacle was solved the LiDAR was mosaicked and separate raster files were created for each town. The LiDAR from each town was reclassified based on the values computed in the table above. Then the rasters were converted to polygons so that they could be easily manipulated with the other vector files. A separate vector file was created for each new sea level rise, which included the new inundation as well as all of the lower inundations. Creating this bathtub model was time consuming and it will be very helpful for future studies. I focused on road and building point inundation. For the road and network analysis I selected the roads in the three towns from E911 roads for the state. I kept the towns connected so that I could analyze how inundation would affect travel between towns. The roads were intersected with each sea level rise to see which roads were directly impacted. Then, using the Network Analysis toolbar, roadblocks were manually entered at each inundated section. Roads cutoff from the rest of town were highlighted and a layer of affected roads was created in addition to the inundated layer. One of the redflags highlighted by this network analysis was the travel time to Midcoast hospital, in Brunswick, Maine. This hospital serves Bath, Topsham, and Bowdoinham, so the shortest distance to the hospital was calculated at each sea level rise using Network Analysis. Finally the buildings inundated by each sea level rise were identified. Bath had building footprints for the whole town. Both Topsham and Bowdoinham only had building point data for residential buildings, which limited the accuracy of the analysis. Other members of this study created more extensive interpretation of building inundation; Cam Adams and Liza LePage focused on building footprints and parcel valuation

5 respectively. DISCUSSION A wide variety of data was created with four people working on these three towns. Tables 2-4 are the results of my analysis of inundated buildings and roads. The numbers are indicative of road segments not entire roads, and cut off roads are ones that are unreachable due to inundation on connecting roads. As is visible on the map, road inundation heavily impacts rural areas with fewer roads. These neighborhoods are more likely to be cut off because they rely on main artery roads. Town Total Roads Roads 1ft Roads 2ft Roads 3ft Roads 4ft Roads 6ft Bath Bowdoinham Topsham Table 2. Number of roads directly inundated at each sea level rise in each town. Town Total Roads roads 1ft roads 2ft roads 3ft roads 4ft roads 6ft Bath Bowdoinham Topsham Table 3. Number of roads cut off by inundation at each sea level rise in each town. Town Total 1ft 2ft 3ft 4ft 6ft Bath Bowdoinham Topsham Table 4. Number of residential buildings flooded by each sea level rise in each town. All of the towns rely on Midcoast Hospital in Brunswick, so I analyzed the routes to the hospital from each town using both GIS and Google Maps. The fastest routes to the hospital from Bath and Topsham were not disrupted by sea level rise. However Bowdoinham was dramatically affected because of downtown inundation. Currently it

6 takes 30 minutes to get from northern Bowdoinham, on White Road, to Midcoast Hospital. Below three feet of sea level rise or storm surge the route to the hospital is unaffected. Above 3ft the time to the hospital increases to 35 minutes. If the bridge from Brunswick to Topsham on route 24 is inundated then the time increases to 40 minutes or more, depending on traffic. While these are not large differences, this time could be vital in an emergency. Ensuring that fast routes to the hospital remain open is a red flag for planners. This report will not discuss the results from other parts of the project. However there were some results common to all of the analyses. First, Bath feels the impacts of sea level rise more strongly than Topsham and Bowdoinham; inundation impacts infrastructure at lower levels and the parcels impacted are of greater value. Second there was a dramatic increase in damage in Topsham and Bowdoinham between 2 and 3ft. Even though Maine is planning for 2ft of sea level rise these towns should look at 3ft to be fully prepared. The accuracy of this model depends on the accuracy of the LiDAR DEMs and buoy tidal measurements; both of these elements carry potential error. Gesch (2009) discusses the merits of using LiDAR as a basis for a sea level rise analysis. It enables greater accuracy and predictions of specific impacts. LiDAR based analysis moves past the general classification of at risk areas produced by past, low resolution models. Cam Adam focused on using this accuracy and quality to produce building footprints and building heights to analyze how flooding would impact individual buildings. However LiDAR also limited this study. The DEMs did not cover the full study area and caused problems throughout the analysis because of differences in the projections. The time

7 spent cleaning up the data was taken from time that could have been spent analyzing impacts such as habitat inundation. Tidal measurements were the other crucial element of the bathtub model. The buoys were used to represent the highest annual tide for the entire town even though each represented one point in a highly variable, estuarine environment. The lack of interpolation created regions where the highest annual tide model was too low and didn t match current sea level. Dan Lesser s analysis of shoreland zoning was hindered by these problems and it likely impacted the road and building analysis as well. The bathtub model was a first step and it was not perfect. This study should serve as a starting not an ending point. Colgan (2009) investigated the economic costs of sea level rise in York County. He performed cost-benefit analysis on reinforcing shoreline infrastructure before inundation or cleaning up afterwards. This study provides a first step towards this sort of economic analysis. Town planners will be able to look at the cost of raising and reinforcing roads now or waiting until inundation becomes a problem. All of these possibilities have been opened to the participating towns. The data produced by this study serves as an introduction to sea level rise analysis in Topsham, Bowdoinham and Bath. The towns can build on this data in a variety of ways. First town planners can use this road inundation model to investigate where roads should be raised or fortified. They can groundtruth and compare this result to past inundation events to improve the model. Using both the building inundation and shoreland zoning developed by Dan, each town can also change zoning codes and require more

8 homeowners to have flood insurance. These sorts of inexpensive steps fit into the low hanging fruit classification. This project can also serve as a model for future work with Bowdoin College or other boundary organizations. Camill et al (2012) completed a more extensive, but similar, project with Brunswick and Harpswell, Maine. They worked closely with their community partners to produce results that were helpful to both parties. Both their study and ours demonstrates the importance of increased communication between knowledge producers (colleges, experts, students etc.) and knowledge users (town planners etc). And, while this study was by no means extensive it will help build interest in the communities for further investigation. This sort of short term project can be a stepping-stone which builds momentum for a larger study. In a study on Tel Aviv and Haifa, Israel, Lichter (2012) emphasized the importance of smaller communities taking initiative and preparing for sea level rise. Defining prudent, small scales for analysis is effective and important for breaking down coastal development into manageable projects. This study supported communities who would otherwise have been stuck. CONCLUSION It is difficult to build interest in planning for sea level rise until after it has already become a problem. This study on the infrastructure impacts of sea level rise in Bath, Topsham, and Bowdoinham was a first step in developing resilient shorelines. My portion of the analysis focused on identifying buildings and roads that may be inundated or cut off by inundation. Shoreland zoning, building outlines, and parcel value were also

9 examined. The creation of a bathtub model of inundation and this preliminary data will help build interest and improve future analysis of sea level rise and storm surge in Sagadahoc County. REFERENCES - Camill, Phil, Maryellen Hearn, Krista Bahm, Eileen Johnson. " Using a boundary organization approach to develop a sea level rise and storm surge impact analysis framework for coastal communities in Maine Journal of Environmental Studies and Sciences. 2,2 (2012), Colgan, Charles; Merrill, Samuel. The Effects of Climate Change on Economic Activity in Maine. Maine Policy Review: 17,2 (2008), Gesh, Dean. Analysis of Lidar Elevation Data for Improved Identification and Delineation of Lands Vulnerable to Sea-Level Rise Journal of Coastal Research: 53 (2009), Lichter, Michal; Felsenstein, Daniel. Assessing the Cost of Sea-Level Rise and Extreme Flooding at the Local Level: A GIS-based Approach. Ocean and Coastal Management: 59 (2012), Slovnisky, Peter, Maine Geologic Survey, 11 November 2012, Personal Communication.