1 of 20

Rock Island partners with Augustana College to comply with Lead Service Line Replacement Act

Erik Bergren, Adriana Reyes, Zack Horve, Kimberly Murphy, Jason Upton, and Peyton Heisch

2 of 20

IEPA #19749

3 of 20

Outline�

Introduction
Data collection and community outreach
GIS predictive modeling
Equitable replacement
Benefits of partnership

4 of 20

Project Leaders

�Michael Reisner, Director, Upper Mississippi Center for Sustainable Communities

Kimberly Murphy, Director, Center for Advancement of Community Health and Wellness

Jenny Arkle, Program Manager, Upper Mississippi Center for Sustainable Communities

Paul Croll, Consultant & Sociologist

Peyton Heisch, Sustainability Interface Manager

Travis Matlick, Water Plant Supervisor

Pat O’Brien, Utilities Maintenance Supervisor

Mike Bartels, Public Works Director

Jason Upton, Utilities Superintendent

Adriana Reyes, Community Data Specialist

Erik Bergren, Equitable Finance Specialist

Zack Horve, GIS Modeling Specialist

5 of 20

Student Speakers�

Single-column body content to go here. Single-column body content to go here. Single-column body content to go here. Single-column body content to go here. Single-column body content to go here. Single-column body content to go here. Single-column body content to go here. Single-column body content to go here. Single-column body content to go here. Single-column body content to go here. Single-column body content to go here. Single-column body content to go here. Single-column body content to go here. Single-column body content to go here. Single-column body content to go here. Single-column body cntent to go here.

Geography, Environmental

Studies, & GIST

Economics & Public Health

Adriana Reyes,

Community Data Specialist

Environmental Studies,

Biology, & GIST

Erik Bergren,

Equitable Finance Specialist

Zack Horve,

GIS Modeling Specialist

6 of 20

Partnership Objectives/Goals

Build Inventory
Propose equitable financing & replacement plans

7 of 20

t�

Inventory Solutions

GIS

Inventory Solutions:

Existing Data Collection

Historical Data
IEPA Monitoring
1986 LCR

Organization

Geographical Information Systems

Community Outreach/ Call to action

News Events
Canvass with sports team
Canvass with College President
Social Media
Community spaces outreach

Strategies

Self Verification Survey
Website with educational and digital survey
Water report link
City Council Ward Representative Outreach
Postcard
Strategic canvassing
Mail out survey
Public Utility Verification

8 of 20

Historical Data�

Compiled and integrated information from > 4000 historical inspection reports

9 of 20

Localized Approach

Website & Postcard

Door-to-door Assistance

10 of 20

Community Outreach / Canvassing�Media day & Football Team

11 of 20

Modeling�Random Forest

12 of 20

Modeling�Results

Image�Space

13 of 20

Modeling�Parameters

Distance from known leads
Home age

Image�Space

A random forest algorithm uses attributes of our verified data to generate a prediction– not lead, galvanized, lead– for each unknown parcel.

Going into our modelling, we did background research and examined the data we collected to identify a wide variety of potentially useful variables.

Through fine tuning the models, we found that the best predictors (lead to most accurate models) of service line composition were a parcel’s distance from verified leads and home age

Distance from verified leads– we found that lead service lines tended to be spatially clustered. (if one house on a block was lead, the likelihood of the other homes being lead tended to increase). Map on the left is a representation of the clustering.

Home age– we also found that the age of the home was important. Homes built during certain ranges of time (particularly early 20th century) were much more likely to have been built with a lead service line.

14 of 20

Modeling�Random Forest

15 of 20

Equitable Replacement

Social Disadvantagement Index (SDI)

Population with less than High School Education
Child Population
Renter Occupied HUs
Population less than 100% Federal Poverty Level (30k)
Per Capita Income

16 of 20

Equitable Replacement

Risk Assessment for Vulnerability of Lead

Children under 5 years

17 of 20

Equitable Replacement

Combination of:

Social & Economic Disadvantagement

Vulnerability for Lead

18 of 20

Replacement Recommendations

Start replacement sooner rather than later and front load replacement schedule

Start with the most disadvantaged blocks

Apply to grant/loan programs as soon as possible

Competitiveness of funding

Complete water service replacements at the block level

19 of 20

Benefits of Partnership

Rock Island Public Works Augustana College

Localized approach
Multi-disciplinary approach
Thorough deliverables

Complete inventory
Prioritization & replacement plan

Real world learning experience
GIS, community engagement, and research skills
Relationship with the broader community
Specific disciplinary knowledge

Public utility insight
Municipal response to laws

Collaborative leadership

20 of 20

THANK YOU!�For Attending This Educational Session!