2003: Climate Impacts
Part 1: Storm and Flood Analysis
Prepare data/background for Impacts and Projections Research Activity
The Impacts Team will conduct an analysis more detailed than that of last summer correlating meteorological parameters of historic storms of the last 40 years in NYC (with themselves, as well as with socioeconomic, financial, and other types of data). This will be done in order to better understand which parameters are associated with storms that are likely to cause the greatest impacts. The team will analyze a minimum of 30 storms (time permitting, preferably 40 storms and at most 50) to produce more quantitative statistics than we were able to produce during the previous summer. Mitch and Marquise agreed on 30 storms as a target after considering time constraints and the definition of a "good" statistical sample most often used in statistical analysis; the "magic number" by the Central Limit Theorem is 30 observations.
The team will explore correlations between the meteorological variables and records of flood heights, and rising sea levels, given by tide gauges. If possible, we should try to separate storm surge from flood height. The flood height is the sum of the storm surge and the tide level--by removing the effects of the tidal cycle we can get a better estimate of the surge, which can be more closely linked to barometric pressure and winds. This can be accomplished by obtaining a data set of the tidal cycle contributions to the surge, then subtracting it from the tide gauge data. Alternatively, we will find, from reputable textbooks or Internet sources, a sinusoidal function that can model this phenomenon to an accurate degree in this area. Using MATLAB (or possibly Excel) we will filter out the effects of the astronomical tide. (We really should not make this an optional activity; it should be mandatory to yield best results). The ultimate objective (although not likely to be completed this summer) will be to connect the tide gauge record of sea level rise, and superimposed storm surges, with changes in storm climatology as projected by Global Circulation Models.
Science Questions:
Which meteorological parameters are most closely associated with storms that produce severe coastal flooding?
How can we use the historical storm data from both meteorological and tide gauge records to make predictions about future coastal flooding?
Data:
6-12 hourly data for: SLP, wind speed and direction, SST, storm tracks.
Mitch is inquiring with the NWS about data availability and Carolyn is looking into extracting these variables from the NCEP data set for the New York area. So far, the availability of this data is uncertain. NWS has not responded as of 5/30/03.
Begin the summer with Adam's statistical analysis of 40 years of storm data: means and anomalies for flood height, as well as a product that depicts storms that are 1, 2, 3, 4, 5 or 6 standard deviations. He will also provide a decadal analysis of these statistics.
Research Tasks
Locate and organize meteorological data (SLP, winds, tracks and SST) in order to analyze along with tide gauge data. Possible sources: NCDC, NCEP, NWS archives. Also, please make sure we have access to the following software: MATLAB, GhostScript and Ghostview (Carolyn, Mitch, Vivien, Marquise)
Gain skills using GIS via tutorials with Spiro and Colin in order to apply in research tasks to understand the socio economic and environmental impacts of storms and flooding on New York City (Marquise)
Cynthia defines model runs that will be used for projections. Determine whether to extract model output for more variables than flood heights. For example, do we want to extract storm characteristics (SLP) to make some assessments about the impacts produced by certain storm types? Adam would like to conduct a historical storm analysis to understand if we can project the amount of flooding a storm with a specific set of characteristics is likely to produce. Identify improvements that the team should seek in revising last summer's projections and specific science questions to address. Model output is compiled and stored on summer workstation for team use. (Cynthia and Rich)
Cynthia/Vivien will define the specifics of the impacts study she would like to the team to conduct. 1) Select metro area regions. 2) Specify impacts to focus on. 3) Identify science questions to address in determining the impacts of storms and flooding in the metropolitan area [these should be in the form of essay questions that the team will respond to explain the impacts and their potential implications]. 4) Identify any specific methodologies that you would like employed.
A team web will be produced throughout the summer in order to provide a research project archive. Adam will be the key responsible team member for developing the web page using electronic forms. Team provides input on design and content.
Train the Team: Review underlying science concepts the team will need to conduct impacts team's research analysis (Topics: sea-level rise, pressure, New York City seasonal weather patterns, ocean-air interactions). (Mitch and Vivien)
- Read and discuss team and individual papers: Identify strengths, and areas for improvement that can be addressed this summer, as well as concepts that students feel need review. (Mitch and Vivien)
- Review and discuss summer goals and products (Mitch and Vivien)
- Review and discuss technical skills and statistical analysis that will be applied. (Marquise)
- Read the Heinz Center Report Human Costs of Coastal Disasters. Discuss ways that this report might be related to Summer 03 Impacts Research (Mitch, Vivien, Marquise, Carolyn [and Cynthia, pending her availability].
Identify a good statistical sample of severe storms from Adam's analysis (30) based on flood heights. This "good" sample could be obtained using the storms having the highest relative standardized ranks with respect to filtered tide heights. Preferably we will choose storms that have high storm surges, as well as low SLPs, etc. In doing this analysis, we must be careful that storms with high (filtered) flood heights, as well as storms with high unfiltered storm surges but low tide heights are included in the sample. If necessary, a random sample of the top 50 storms may be selected to minimize statistical bias. Work with Vivien to determine and document methods used to identify these storms that will be used in the summer research, as well as those applied to Adam's analysis (Marquise)
For each of the 30 storms selected, we will collect this information and input it into a spreadsheet for easy manipulation: SLP, wind speed and direction, track and SST. Next, correlation plots and statistics will be prepared of tide heights versus each meteorological variable. Marquise analyzes 12 storms since he starts earlier than the high school students. Denise, Adam and Domilola each conduct statistical analysis for 6 storms. Marquise will finish his analysis by July 7th and begin to work on the second part of the study dealing with impacts while also mentoring the high school students along with Mitch.
Scientific questions for students to answer in preparing the storm analysis:
- Which storm characteristics are most closely associated with the storms that produce the greatest flooding (in terms of flood heights)?
- Is any one of the variables a more valuable predictor of this flooding?
- Is there a historical pattern in the storms and their correlated characteristics?
Research Products
Plots of correlations between flood heights of 30-40 storms and associated meteorological data. Plots may be condensed for space purposes, numerical correlation coefficients and multiple correlation/regression analysis may be used in areas of project requiring objectivity (provided that it is not too complex) to make the study more quantitative. I think the correlation coefficients themselves are more valuable than the plots for this purpose.
Team members will produce a written assessment of correlations. We will identify questions arising from the results, and attempt to explain these questions and the implications of the team's results for future storms in the New York metropolitan area and their potential impacts.
Model runs and output for projections
Identification of relevance of Heinz report to ICP Impacts' team research and specific contributions that the Impacts team can make to the Heinz study
Part 2: Impacts Study
This part of the study focuses on using socioeconomic data from the U.S. Census to qualitatively and quantitatively assess the potential impacts storm damage in the New York Metropolitan area. At the end of the study we hope to understand the social and economic costs that affect coastal populations at risk in this regard. One key goal is to be able to advise policy-makers of these populations and make recommendations that may result in lower human costs. Ideally 2-3 contrasting neighborhoods will be assessed (starting with ones already used in previous studies such as Rockaway Beach, Jamaica Bay/Broad Channel, and Lower Manhattan). The aim is to refine summer 2003 analysis to greater detail and produce overlays of socio-economic and environmental data, in order to assess the potential economic impacts of predicted coastal flooding.
In conducting this analysis we will attempt to apply economic analysis techniques. Toward this end we have discussed Marquise getting advice from one of his Economics Professors at Cornell, as well as Cynthia introducing us to a colleague. Unfortunately Marquise was unable to make this contact at Cornell. Once the project is really flushed out, we will decide whether we need to ask Cynthia to make the connection for us to her contact.
Data:
U.S. Census data for Population, Population Density, Income, Age, Race. We may also be able to find information about housing type and value for different metropolitan areas.
Economic Data: To assess the costs of storm damage we should check with the Cloud team for the financial data set they used in Summer 03. Also Mitch is checking with NWS to see if such a data set exists.
Science Questions
What are the human costs of storm damages?
What is the range of economic costs of the sample of storms studied?
Are their specific populations (i.e. racial or income groups) at greater risk of incurring impacts than other?
Research Tasks:
Extract and organize the socio-economic data for analysis
Identify elevations of areas proposed for impacts study
Confirm with Spiro in Cynthia's group as to whether we can use Mitch's 365-day version of ArcView explorer. In addition, Mitch has a copy of the following paper, using GIS, which may be a useful starting point: http://www.brown.edu/Students/SFT/warwick/index.html.
Use GIS to produce analysis maps of impacts of storms and flooding: 1) characteristics of population impacted [age, income, population density, etc, 2) economic impacts [housing costs, businesses at risk], social costs [health], and environmental [coastal damage].
Research Products
Integrated maps of the storms and associated characteristics and socioeconomic data for the area studied.
Essays that address team's assessment of the impacts of storms and flooding.
Part 3. Projections
The final stage of the project overlaps with the impacts study and will involve the team in making projections for the future impacts of storms and flooding in the New York City metro area. Model data will be used from 50-year runs to project the flood heights and possibly select storm characteristics. These model projections will be integrated with the with the impacts study to discuss how the New York City metro area may be impacted in the future by storms and flooding.
Data: 50-year model run data (which model???). Projected flood heights
Research Tasks
Integrate projected flooding (and possible storm characteristics, like SLP) into impacts GIS mapping analysis
Develop and apply methodologies with Cynthia and Vivien to assess future impacts
Prepare responses to essays provided by Cynthia and Vivien to write assessment
Research Products
Team addresses a series of essay questions provided by Cynthia and Vivien in order to explain the future impacts of storms and flooding on the New York Metro area.
Team web page with all the data, results and written analysis
PowerPoint Presentation (5-10 viewgraphs) for the informal final team presentation that we will be holding on August 1st.
Adam, Denise and Domilola produce an Intel Papers:
Historical Record of NYC metro area storms
- What are the historical trends in storms in terms of their intensity (SLP), frequency and associated flood heights that impact the New York metro area? Consider current conditions and lessons from the past.
- Can the most severe storms be described by a set of common set of characteristics (winds, SLP, flooding, and location, i.e. coastal/in-land)?
- Plot: Historical trends line of SLP versus storm surge
- Histogram of daily high tide heights and identify storms that resulted in tide gauge readings above 2.5 SDEV with respect to the sample of daily high tide heights.
- Table that identifies date ranges associated with each storms and indicates storms 2.5 and above
- Table of SLP, surge and winds for storms that have hit NYC metro area (1960-2000)
- Correlations of storm characteristics (SLP, winds, track) and the storms that result in severe flooding.
Case study of the vulnerability of Jamaica Bay, Rockaways and Lower Manhattan
- How might storms impact in these areas?
- What is the vulnerability (high to low risk) of these areas? Describe in terms of potential impacts on demographic, socio-economic and ecosystems?
- Table of the following demographic and socio-econonmic data from the US Census. Population by race, age, density, median income, people living below poverty, housing value, buildings built before 1940
- Descriptive characteristics about the potential ecosystems impacted
- Description of trends of ecosystem, infrastructure (buildings, roads, water and energy systems), industry, employment and demographics characteristics.
- Identification of population and infrastructure at risk.
- Design a vulnerability index (time permitting)
- GIS plots of each region that define the relevant demographic and socio-economic characteristics for the impacts study.
Climate Projections
What are the future potential impacts of storms and flooding in Jamaica Bay, Far Rockaway and Lower Manhattan?
- Incorporate model projections (2020, 2050, 2080) of flooding into the GIS maps
- If possible, define the standard of error in the GCM projections and produce a range of potential future flooding
- Assessment of projected future flooding on regions of interest based on GIS information
- Apply Adam's storm methodology (provides a tool to assess the types of storms associated with the projected flooding) to compare historical storm damage with comparable projected storms and assess their potential damage.
