COVID-19 Online Training
March 15, 2020
Machine Learning and Analytics Modeling for USACE
February 9, 2020

Weld County is located in Colorado and has an area of over 4,000 square miles and a population over 325,000 residents. The county seat is Greeley and comprises the Greeley, CO, Metropolitan Statistical Area, which is also included in the Denver–Aurora, CO, Combined Statistical Area.


iParametrics was retained to provide data science and analytical analysis in support of updating Weld County’s 2016 Multi-Jurisdictional Hazard Mitigation Plan in collaboration with local municipalities, fire departments, school districts, and other community organizations. This plan is designed to reduce the risks posed by hazards that affect the county and must be updated and approved by FEMA every five years to keep it current and to maintain eligibility for certain types of disaster assistance. Hazard mitigation planning helps residents, business owners, elected officials, and municipal departments think through how to plan, design, build, and establish programs for risk reduction.


We were tasked with developing an analytical and quantitative approach to answer the question: Where are the areas of highest risk within Weld County? The project aimed to incorporate advanced analytics and geospatial modeling into the Hazard Mitigation Plan (HMP) for Weld County by developing a geospatial analytics model to execute the risk identification component of the HMP.


The risk analysis for the Weld County HMP was performed using an analytical suitability model that was created in the GIS software, Esri ArcGIS. In general, a suitability model was used to identify the most fitting areas based on specified criteria and data and was employed to answer the question of where the highest risks can be found. Criteria data was curated, prepared, and used for each hazard type of interest. The hazard types, corresponding datasets, preparation methods, and resolution were all customized based on the project requirements and need outlined by the client. To combine the various hazard-specific layers into a composite risk layer, each data layer was transformed to a common hazard format and scale. County analysis and categorical data was transformed using the Unique Categories transformation method while continuous data was transformed using the Linear option under the Continuous Functions transformation method. In addition to handling the transformations, the model also allows for control over the criteria weight before generating the composite risk scores and mapping. For this analysis, the weights were changed to reflect the potential magnitude of the population and spacial features that would be affected in the case of the corresponding hazard event. After establishing the weights, the model was used to run the calculations and output the composite risk analysis for inclusion in the hazard mitigation plan.


iParametrics developed a suitability model in Esri ArcGIS Pro to identify the areas of highest risk to Weld County. The model ingested over one hundred data sources from over 10 different state and federal agencies to represent the hazard types of highest concern to Weld County, such as severe storm, prairie fire, hazardous material, crime and public health. In addition to gathering the data for the model, we also transformed the raw data, which varied in type (point, line, and polygon vector data and raster data) into continuous raster layers using various geoprocessing tools, such as distance, spatial join, and geocoding

The hazard types were then weighted relative to one another based on impact analysis and combined the various hazard layers to generate a composite risk layer. From the model, we were able to provide the County with hazard-specific maps detailing the data sources used for each hazard type, as well as a composite risk map for the County. The approach offered many advantages, including:

  • Objective, data-driven outcomes
  • Flexible and open methods
  • Reproducible workflows and results
  • Easily updated analyses over time and after events
  • Defensible output

The analysis was recognized by the State Hazard Mitigation Officer as a possible new standard for hazard mitigation planning.