Predicting parcel scale redevelopment within the Berkeley neighborhood in Denver, Colorado using linear and logistic regression

Abstract

Many water resource issues associated with urban development result from increased impervious cover. As impervious cover increases, rainwater infiltration decreases leading to increased flows and potentially higher pollutant loads in the runoff. Most of the prior research on this topic investigates the increase of impervious cover through the transformation of undeveloped to developed regions, or the expansion of dense urban development into outlying suburban areas. A topic that is not as widely understood is the impact of infill redevelopment on stormwater runoff. Infill redevelopment is rapidly occurring in many Denver neighborhoods, where previously developed properties with low-density structures are being replaced by larger, higher density units. Regardless of impervious cover increase, these redevelopment projects are only required to incorporate stormwater detention and treatment systems if they are greater than one acre. Due to most of the redevelopment in Denver (86%) occurring on sites less than one acre, the burden of stormwater treatment and detention ultimately falls on the city. This study focuses on modeling the spatial distribution of infill re-development on a parcel scale to investigate its cumulative impacts on stormwater quality and quantity for near-term and future conditions. Future redevelopment and imperviousness is determined by distributing a “business as usual” linear growth scenario to the parcels with the greatest probability of future redevelopment. Then, a logistic regression model is used to determine the parcels that will be redeveloped. Results indicate that building cover change within study site from 2004 – 2014 followed a linear pattern. During this period the total building cover increased by 17% or 1.7 % per year on average. The logistic regression model determined the total value, year built, percent difference between current and max building cover, the current use classification: rowhomes, and current use classification apartments to be the greatest predictors of redevelopment, resulting in a model that was 81 % accurate. The "Building to Land Area Ratio" variable was found to be highly correlated with the “Improvement to Land Value Ratio”. However, the “Building to Land Area Ratio” was found to be a better predictor of redevelopment. The final model estimated an increase of 820,498 sq. ft. (18.8 acres) in building coverage between 2014 and 2024. This method will provide municipalities with a tool that can be used to estimate parcel scale impervious cover growth from publicly available planning data resulting in more informed urban watershed planning and policy development.