Projects

Determining best practices when reseeding neighborhoods with nonprofit rebuilding after coastal storms

End Date: 1-31-2014

Abstract

The Phoenix of New Orleans (PNOLA) arose in Mid-City in response to community needs after Katrina, and was one of the first nonprofit organizations to return to New Orleans. The Mid-City district is unique in that PNOLA is the sole nonprofit organization directly assisting with home rebuilding; this has made it possible to analyze the effects of one charitable organization with long-term presence in a predefined spatial extent. We have established a community-university partnership between PNOLA, the University of South Alabama, and the University of Central Arkansas that will combine PNOLA’s records of Mid-City rebuilding with spatial GIS layers of elevation, flooding, and city building permits to gain insight into the recovery process in Mid-City. We hypothesize that the houses rebuilt with a local nonprofit organization’s assistance, “seed” houses, resulted in positive spatial spillover effects to neighboring houses that did not receive assistance from PNOLA, and that the spatial distribution of PNOLA-assisted homes influenced the properties of their spillover effects. Our objectives are to determine whether spillover effects occurred, using exploratory spatial analysis and dynamic spatial autoregressive probit models. If so, we plan to quantify those effects in relation to distance from a PNOLA-assisted home or homes. Our overall goal is to determine the optimal spatial distribution and cluster size of seed houses to maximize spillover effects; this knowledge will assist planners and decision-makers in allocating resources to future disaster recovery efforts. Researchers, planners, nonprofit groups, and residents alike will want to know how rebuilding has progressed both spatially and over time, and how the resources of external groups can be used efficiently to speed rebuilding and recovery.

Objectives

We hypothesize that the houses rebuilt with a local nonprofit organization’s assistance, “seed” houses, resulted in positive spatial spillover effects to neighboring houses that did not receive assistance from the Phoenix of New Orleans (PNOLA), and that the spatial distribution of PNOLA-assisted homes influenced the properties of their spillover effects. Our overall goal is to determine the optimal spatial distribution and cluster size of seed houses to maximize spillover effects; this knowledge will assist planners and decision-makers in future disaster recovery efforts.

Our objectives are to determine whether spillover effects occurred, and if so, to quantify those effects in relation to distance from a PNOLA-assisted home or homes: 1) to determine whether residences rebuilt with PNOLA assistance affected the likelihood or timing of home rebuilding in the immediate surroundings, 1a) to test the size of the area affected by a “seed” house, and 2) to determine whether the spatial distribution of PNOLA-assisted residences influenced their spillover effects.

Methodology

In all models, our dependent variable will be binary, whether a home had begun to rebuild (1) or not (0) at each timepoint. Our explanatory variable will be the distance from each home to the nearest PNOLA-assisted home or homes. Covariates to be included will address confounding spatial effects, such as  distance to the nearest rebuilt non-PNOLA home and the nearest operating school, as well as confounding temporal effects such as the timing of infrastructure and service return to the neighborhood.

First, we plan to use Geographically Weighted Regression (GWR) in exploratory spatial analysis. GWR allows the exploration of local variation in the relationships between the predictor and outcome variables. This would allow us to test, for example, whether the same relationship between the predictor variables and rebuilding exists near the Esplanade Ridge (a strip of higher elevation that runs along Esplanade Avenue) as it does in other parts of Mid-City. GWR will allow us to easily compare coefficients and t-values of explanatory variables in different zones of Mid-City.

Second, we plan to use a dynamic spatially autoregressive (DSAR) probit statistical model to achieve our objectives. DSAR probit can account for spatial autocorrelation as well as the non-normal distribution of rebuilding data coded as binary values. DSAR probit is well suited to analyzing changes over time and can flexibly accommodate complex model structures using Bayesian methods of estimation.

Rationale

The Phoenix of New Orleans (PNOLA) arose in Mid-City in response to community needs after Katrina, and was one of the first nonprofit organizations to return to New Orleans. PNOLA has gutted 200 properties, completely or partially rebuilt 73 homes, and plans to rebuild 45 homes in 2011. The Mid-City district is unique in that PNOLA is the sole nonprofit organization directly assisting with home rebuilding; this has made it possible to analyze the effects of one charitable organization with long-term presence in a predefined spatial extent.

PNOLA has kept continuous records of rebuilding progress since 2005, for homes it has assisted and for all residences in the Mid-City district (the area bordered by Canal, Broad, Tulane and Claiborne). Thus, every residential address in this area has been monitored for outward signs of occupancy, repair, and rebuilding for the past five years. Dr. David T. Mitchell has established a community-university partnership between PNOLA, the University of South Alabama, and the University of Central Arkansas that will combine PNOLA’s records of Mid-City rebuilding with spatial GIS layers of
elevation, flooding, and city building permits to gain insight into the recovery process in Mid-City. This project will add to the literature on evacuee returns because we have measures of exogenous rebuilding by a nonprofit organization.

Because nonprofits have finite resources to work with, and because large-scale coastal storms often tap multiple funding sources, it is imperative that recovery funds are allocated wisely. The natural experiment of post-Katrina rebuilding, combined with PNOLA’s consistent long-term monitoring of Mid-City’s recovery, offers a way to determine whether the homes rebuilt by PNOLA had a positive neighborhood effect on the surrounding area, and if so, what spatial pattern of “seeding” houses is more effective. We expect that the results will assist local governments, nongovernmental organizations, and community residents in planning for optimal resource allocation for rebuilding after future coastal storms.

For More Information Contact: the MASGC Research Coordinator, Loretta Leist (Loretta.leist@usm.edu). Please reference the project number R/MG/CSP-23.