Large Culvert Network Resilience Metric


New York State Department of Transportation’s (NYSDOT) large culvert network has experienced an increased rate of failure during severe storm events. This is linked in part due to the well documented increase in the intensity and frequency of such events.

NYSDOT has a well-developed inventory and inspection system for large culverts, but lacks a network-wide assessment of the hydraulic capacity or resilience of this system. The intent of this project is to develop that assessment, with an individual resilience metric of some type for each such structure in the NYSDOT network. This may take the form of a hydraulic demand/capacity ratio, return period for passable flows or some other suitable metric. The purpose of this is to provide a new tool aide in better prioritizing capital program investments in a manner that improves system resilience at both a site and network level.

Research Objectives

The purpose of this study is to develop a system-level assessment of NYSDOT culvert infrastructure resiliency. To achieve this, the research team will do a thorough literature review, summarizing the methods that are available to calculate culvert resiliency. They will then develop hydrological models for the wide range of catchments that are served by these culverts. To achieve this, the team will delineate catchment areas using high-resolution, publicly available digital elevation models. Finally, the parameters of conceptual models will be regionalized for the entire culvert network. Precipitation intensities corresponding to various return periods will be used to analyze the exceedance probabilities of culvert flows. The primary outcome of this research will be a tabular and map-based quantitative assessment of the culvert infrastructure resilience.

The proposed research objectives shall be achieved through the following tasks.

Task 1: Project Management

Upon starting the project, the research team members will meet with the NYSDOT Project Manager to present the planned schedule and technical activities. Meeting notes from the kick-off meeting will be submitted by C2SMART to the NYSDOT Project Manager.

The team will also submit one-page Quarterly Reports in Microsoft Word format by the tenth day of the first month of each calendar quarter (January 10, April 10, July 10, and October 10). The reports will include tasks accomplished during the preceding three months. 


• Summary of kick-off meeting 

• Quarterly SPR status reports

• Project briefing

Task 2: Literature review

The research team is familiar with past work and recent developments on various frameworks of rainfall-runoff estimation for pipe design. However, to ensure that the proposed project follows the state-of-the-art research on the estimation of hydraulic demand of a large-scale culvert network, the team will first carry out an extensive literature review. 

Based on comprehensive analysis of literature review and researchers’ knowledge and experience in this area, a summary of current state-of-the-art will be prepared and a detailed presentation on the outcome of the literature review will be made to the NYSDOT Technical Working Group (TWG) for the project.

Deliverable: A report on literature review and a PowerPoint presentation to the TWG

Task 3: NYSDOT culvert data preprocessing 

The NYSDOT inventory data does not have full geometry information about the height of the culvert opening.  However, this information can be found in the inspection reports. The team will compile this information for culverts from the available inspection reports. For culverts where such information is not available, they will employ heuristics such as that any arch has a rise equal to 2/3 of the span.  

Deliverable: Upgraded culvert database will the dimensional information required to estimate hydraulic capacity

Task 4: Develop catchment areas for each large culvert in the NYSDOT system

In this task, the research team will use the inventory data from NYSDOT to geolocate all the culverts within the New York State within a GIS environment. They will then download high-resolution topographic data that is publicly available to do catchment delineation and classification. The team will also evaluate the adequacy of already-available catchment delineations in open-access scientific repositories like HYDROSHARE.    

Estimation of catchment area is important for calculating the total rainfall collected by the catchment which finally turns into runoff. The team will also tabulate other morphometric characteristics like average slope and land cover for each catchment, which will in turn assist with the more representative parameterization of the conceptual hydrologic models. 

Deliverable: Tabulated information on the catchment area, average slope, and other essential morphometric characteristics of the catchments served by NYSDOT culvert network  

Task 5: Develop hydrologic models

Based on the catchment delineation and characteristics, the team will ascertain the differences and benefits as provided by the usage of the rational method of runoff calculation, NRCS curve number method, and a nonlinear reservoir based ordinary differential equation model. These hydrologic models will then be run using precipitation time-series corresponding to various duration and various intensities. This information will be collected from the publicly available precipitation data and corresponding intensity duration frequency curves. 

Deliverable: Model(s) agreed upon between the team and the TWG, with specified parameter values, for all the catchment of the culvert network

Task 6: Resilience metric for each culvert in the system

Subtask 6a: Using the designed hydrologic models, stormwater discharge corresponding to 2-, 5-, 10-, 25-, 50-, and 100-year rainfall return periods will be calculated employing county-specific in-tensity-duration-frequency curves and using rainfall intensities corresponding to the 1-h to 24-h duration. This analysis will be used to calculate hydraulic capacity of all NYSDOT culverts based on the provided dataset with consideration of culvert inlet headwater depth impacts.

Subtask 6b: NYSDOT’s hydraulic analysis performed using the HY-8 tool for recently designed and built large culverts will be used to compare and validate NYU’s hydraulic analysis.  This com-parison will provide quantitative estimates related to the agreement between the developed hydro-logical models and the methods used by NYSDOT during design. Any mismatches can be further used to tune and calibrate our methods, which will be deployed for all the culverts. 

NYSDOT has some (undefined) number of more recent hydraulic analysis results for recently de-signed and built large culverts. These are not contained in any database, and are instead part of the individual project files. It is expected that many/most of these would have been performed using the HY-8 tool. These are cases in which there should be good data for capacity and expected demands/flows. NYSDOT will begin an effort to collect information on such structures and this will help with validation of hydrologic models developed by NYU team. 


• The return period of the passable discharge for the entire NYSDOT infrastructure reported as a resilience metric (e.g., demand to design ratio). 

• System-, region-, and catchment-level insights on the adequacy and resilience of culvert infrastructure summarized in a resilience and risk map. This mapping will be broken down into at least twelve areas, including one map for each of the eleven NYSDOT regions and one statewide map.  

Task 7: Uncertainty analysis for the hydrologic models 

Hydrologic systems are complex constituting of spatially distributed and temporally variable properties and processes. Such systems are generally modeled mathematically after some simplifying assumptions, for example, by aggregating certain variable properties and neglecting many subprocesses. These model structure deficits result in a mismatch between the real and the modeled system response, which entails uncertain predictions. In addition to model structure deficits, the presence of errors in model inputs and in the observations of system response contributes to the uncertainty in the estimation of model parameters. For risk-based decision making using these models, their uncertainties need to be adequately and explicitly stated. Therefore Monte Carlo simulations will be run using a range of parameter values and analyze the effect on the hydrologic models. This uncertainty analysis will help to determine impacts on the overall network resiliency characterization. 


• An assessment of the sensitivity of the evaluated system resilience to various sources of uncertainty. 

• A statistical tool, as an add-on to the hydrologic models, that evaluates the 90% confidence interval of the model output.

Task 8: Scenario analysis using updated intensity-duration-frequency curves to account for potential effects of climate change

Given the current and anticipated changes in precipitation patterns globally, and especially across the United States, the adequacy of stormwater infrastructure is contingent of the nature of intensity, duration, and frequency of precipitation events. Several studies have confirmed increases in both the frequency and magnitude of extreme precipitation in the northeastern U.S over the past decades. According to the Intergovernmental Panel on Climate Change (IPCC), the frequency and magni-tude of extreme precipitation in this region will likely continue to increase. To quantify the potential effect of climate change on the resilience of the NYSDOT infrastructure, analysis will be repeated using IDF curves corresponding to RCP 8.5 and RCP 4.5. For this analysis publicly available rain-fall projections from the Northeast Regional Climate Center Cornell University, Ithaca, NY, will be used.

Deliverable: Insights on the adequacy and resilience of culvert infrastructure summarized in a re-silience metric and risk map under projected precipitation patterns. This mapping will be broken down into at least twelve areas, including one map for each of the eleven NYSDOT regions and one statewide map.  

Task 9: Final Report

The researcher team will submit a final report summarizing the work done in Tasks 2 thru 9, with special attention given to Task 5. In the context of this task, we will deliver completed resilience metric for each culvert in the system – and include both tabular data set and graphic map representations.

Deliverables: Final Report, Technical Briefing


Related Media


Omar Wani

Assistant Professor, NYU

Omar Wani is the Principal Investigator on this project.

Frank DarConte

Research Professor, NYU

Frank DarConte is a Co-Principal Investigator on this project.

Hani Nassif

Associate Director, C2SMART
Professor, Rutgers

Hani Nassif is a Senior Collaborator/Advisor on this project.

Omid Emamjomehzadeh

Student Researcher, NYU

Omid Emamjomehzadeh is a Student Researcher on this project.