Land Use Features Driving Human Health: A Zip Code Analysis in California

Abstract

Land cover has been linked to local health effects both positive and negative (Patz 2004; Kamp & Davies 2008; Holgate et al., 1999; Stansfelt, 2000; Mace et al., 2004). For example, aircraft and urban noise-scapes altered from natural land cover have been linked to mental health issues and cardiovascular disease along with elevated levels of stress (Kryter, 2009; Holgate, 1999; Kamp & Davies, 2008). Mining and industrial coal factories also deviating from natural land cover have been linked to increased mortality (Hendryx, 2016; Epstein, 2011; Buonocore, 2016). Conversely, community access to green-space, parks, and higher biodiversity areas appear to be linked to increased health (Velarde, 2009; Pretty, 2005; Karjalainen, 2010; Ostfeld, 2006). However, research is lacking to map a correlation between local human disease incidence and vegetative land cover. Land cover data can be compared as samples at the zip code grain via GIS government databases and correlated via regression analysis to local health markers of the common diseases. The central research question I addressed was: Can land cover type be correlated to human health indicators? As an interesting case study example to be included also in the analysis, the EPA List of Toxic Waste sites was also be regressed against cancer and heart disease to see if a high proportion of land cover that have been highly altered from natural land cover drives poor health. To model if natural land cover promotes or lessens human health, samples of zip codes in California State were used from government data. The regression analysis accounted for confounding factors of gender, income, race, education, employment, and access to healthcare, which are the factors that are also expected to affect health (WHO, 2019). A multiple regression analysis was performed to examine Statistical correlations between land cover and health by using a proxy of the two most common diseases, cardiovascular diseases and cancers. Based on previous research, the expected hypothesis was that more vegetative land cover will correlate to decreased local cardiovascular disease and cancer disease while abundant land cover types such as forests and vegetated areas will correlate to increased health. This predicted correlation between local land cover and human health was not statistically supported, perhaps due to multicollinearity showing among variables and missing variables in the equation. More research might be needed to more rigorously test the hypothesis. The second hypothesis that a higher incidence of waste sites and toxic sites within a zip code area will correlate with decreased human health in the form of more cancer and cardiovascular disease was also not able to be statistically supported, probably also due to multicollinearity and additional variables were missing from the equation. In the future, research of this type may be useful so that social or regulatory pressure could guide taxes or penalties placed on new developments of land that affect community health via altering natural types of land cover or creating waste sites.