HEC-FDA, Version 2.0

The model certification plan defines the scope, schedule, and budget to review and certify HEC-FDA, Version 2.0, for use in U.S. Army Corps of Engineers (USACE) planning studies. The Flood Risk Management Planning Center of Expertise (FRM-PCX) is responsible for review and certification of HEC-FDA in accordance with Engineering Circular (EC) 1105-2-412, as amended. The HEC-FDA software, hereafter referred to as software, was developed by the USACE Hydrologic Engineering Center (HEC), the software proponent. The technical quality, system quality, and usability of the software will be evaluated, as well as its conformance with current USACE policy.

Background

Purpose of Model

The primary purpose of the HEC-FDA software is to assess flood risk. The assessment of flood risk as measured in expected annual damage is used to quantify the risk of flooding in a study area and evaluate the potential benefits achievable through alternative flood risk management plans. Project performance statistics are calculated as part of the risk assessment. The project performance statistics serve the secondary purpose of completing an evaluation of the National Flood Insurance Program (NFIP) Levee Accreditation.

Software Description and Depiction

The HEC-FDA software implements the risk assessment methods discussed in Engineering Manual (EM) 1110-2-1619 and draws on those methods to produce the results required for compliance with Engineering Regulation (ER) 1105-2-101. Monte Carlo sampling techniques are the driving method of including knowledge uncertainty and natural variability in the assessment of flood risk with uncertainty. Please note that knowledge uncertainty and natural variability are not evaluated separately for most input relationships. Often, knowledge uncertainty and natural variability are wrapped into a single characterization of uncertainty, or natural variability is the sole factor by which the Monte Carlo simulation is implemented. An analytical flow-exceedance probability function is an example that includes both sources of uncertainty separately in estimating the uncertainty about the function but which result in a single distribution for each coordinate. Natural variability is reflected by the Log Pearson Type III probability distribution of flow, meanwhile the equivalent record length is used to reflect the knowledge uncertainty about the true series of annual peak flows. Uncertainty around a structure's value may represent one or the other type of uncertainty depending on the approach used by the user to quantify the uncertainty about the structure value.

Contribution to Planning Effort

USACE Civil Works Planning requires the use of risk analysis procedures for formulating and evaluating flood risk management measures. Pursuant to the Flood Control Act of 1936, flood risk management projects must be economically justified to be authorized and constructed. As such, the assessment of risk in the with- and without-project condition is a required step to assess the potential reduction in flood risk (i.e., the benefits) obtainable through flood risk management measures, while considering the probabilistic nature of storm damage and the uncertainty in the measurement of input data and modeling.  Benefits are derived by comparing the average annual equivalent damage when a flood damage risk management measure is in place (the with-project condition) with the average annual equivalent damage that would occur in the absence of any project (the without-project condition). HEC-FDA is intended to provide users in the USACE Planning community with a standard analytical tool to calculate flood damages and benefits under these conditions.

Description of Input Data

There are roughly four categories of input data: hydrologic engineering, hydraulic engineering, geotechnical engineering, and economics. The hydrologic engineering inputs consist of a flow-exceedance probability function with uncertainty and optionally an inflow-outflow relationship. The hydraulic inputs consist of hydraulic modeling for a series of 8 hydraulic events, a rating curve with uncertainty, and optionally an exterior-interior relationship with uncertainty. Optionally, geotechnical engineering is included to represent lateral structure specifications together with a fragility function. Finally, the economic data consists of a structure inventory, depth-percent damage relationships, and uncertainty parameters specific to occupancy types which are used together to produce an aggregated stage-damage relationship with uncertainty. This input data is modified to reflect different with-project conditions, such as the presence of a levee or nonstructural actions. For more information on this data, please see Engineering Manual (EM) 1110-2-1619.

Description of Output Data

During each iteration of the Monte Carlo simulation, flood magnitude and flood damage realizations are saved to the database. A converged simulation produces a distribution of expected annual damage, a distribution of average annual equivalent damage, a distribution of annual exceedance probabilities, and a distribution of event-specific stages. These distributions are queried to produce the results required by Engineering Regulation (ER) 1105-2-101 and Engineering Regulation 1105-2-100. For more information on the output data, see HEC-FDA User Documentation for a list of essential HEC-FDA software resources.

Statement on the Capabilities of the Software

HEC-FDA is designed to simulate flood risk over a user-identified period of analysis in a particular study area. The primary capabilities include a graphical user interface, Monte Carlo sampling techniques, estimation of expected annual damage and project performance statistics, and graphics and reporting.

Graphical User Interface

Users interact with HEC-FDA through a graphical user interface. The main focus in the design of the interface seen in the image above was to make it easy to use the software while still maintaining a high level of efficiency for the user. The interface provides for the following functions:

• File management
• Data entry and editing
• Expected and average annual equivalent (AAEQ) damage analysis
• Project performance analysis
• Tabulation and graphical displays of input and output data
• Geographic information system (GIS) processing using HEC-RasMapper
• Reporting facilities
• Context sensitive help

Monte Carlo Sampling Techniques

HEC-FDA applies both natural variability and knowledge uncertainty through Monte Carlo sampling. Each iteration in a simulation represents a realization of input relationships and coinciding flood risk that could exist given the data uncertainties in the model. The results of the analysis provide a distribution of estimated consequences from a hazard curve. Below is a list of inputs which include uncertainty in their input and application. 

• Flow-frequency relationship
• Stage-frequency relationship
• Stage-discharge relationship
• Exterior-interior relationship
• Aggregated stage-damage relationship
  • First floor elevation
  • Structure value
  • Content value
  • Vehicle value
  • Depth-percent damage
• Geotechnical fragility function

Graphics and Reporting

Input and output data are available through plots, tables, and GIS data layers. The expected annual damage (EAD), average annual equivalent (AAEQ) damage and annual exceedance probability (AEP) are provided in a tabular summary and plotted in a histogram. Other project performance statistics are provided in tabular summaries. Examples of this reporting can be seen in the images below.

Description of Software Development Process

The software development process workflow followed standard DevOps procedures, including a robust branching strategy. Review the HEC-FDA Contributing Guide to learn more about team development procedures. Changes to the software codebase are tested against several hundred unit tests. HEC-FDA underwent HEC's internal review process prior to release. The process consists of a formal review of both the software and user documentation. The process also includes evaluating the results of existing studies, including London-Orleans, West Sacramento, Folsom Dam Raise, River Des Peres, Sabine to Galveston - Freeport, and Muncie, Indiana. Any comments or concerns are addressed prior to release.

Technical Quality

Theory

The theory underpinning the methods used in HEC-FDA are described in detail in Engineering Manual 1110-2-1619. In short, the flow of a river is connected with the elevation of the flow and the damage from the elevation of flooding. The probabilistic approach produces a distribution of expected annual damage among other results. The HEC-FDA model wiki on GitHub describes the methods, procedural steps, and computations that HEC-FDA uses to compute the the results with uncertainty. The wiki describes each input relationship and any underlying parameters, how the relationships and parameters are used, how they influence results, and where they come from.

Description of the System Represented by Modeling

The configuration of study data in HEC-FDA results in a model (a concrete representation of a system) of flood risk. The magnitude of the hazard is connected with the chance that the hazard enters the floodplain, a measurement of the damageable assets in harm's way, how vulnerable to the hazard are the damageable assets, and the consequence of the hazard reaching the damageable assets. The hazard is represented by a flow- or stage-exceedance probability function, an inflow-outflow function, a rating curve, and a spatially-referenced data set of the magnitude of flooding for 8 unique events. The chance that the hazard enters the floodplain is a function of the geotechnical performance (a levee fragility function). The value of damageable assets is recorded in depreciated replacement value for structure, contents, and vehicles. The vulnerability of the damageable assets is determined using the height of the first floor above ground and a relation between the depth of flooding and a depth-percent damage relationship, a relation which produces an estimate of the consequences. The linking of these objects results in a realization of a damage-frequency function - a function that is integrated to compute a realization of expected annual damage.

Analytical Requirements

The analytical requirements necessary for a functioning and complete HEC-FDA model are described in the Introduction of the HEC-FDA User's Manual. These requirements include 2-Dimensional (2D) hydraulic modeling, a flow-frequency function and rating curve or a stage-frequency function, a fully-specified structure inventory, and a collection of depth-percent damage functions.

Assumptions and Limitations

• Uncertainty about hydrologic and hydraulic input data is computed external of HEC-FDA.
• HEC-FDA is not based on probabilistic life-cycle analysis. This means that evolution of the system in response to system events is not captured. There is no concept of time in an HEC-FDA model. The frequency-based approach involves a series of iterations that sample from the same distributions of input relationships each iteration. This approach results in a number of limitations including most importantly the inability to model the transfer or transformation of risk.
• The only hazard used in the estimation of consequences is the depth of flooding.
• Expected annual damage from the most-likely future year until the end of the period of analysis is held constant.

Conformance With USACE Policies and Procedures

HEC-FDA implements the methods and procedures presented in Engineering Manual 1110-2-1619. Risk and uncertainty is included in the modeling in conformance with Engineering Manual 1110-2-1619 and Engineering Regulation 1105-2-100. A parametric bootstrap method is used to produce a distribution of flows for a given frequency that approximates the distribution that would be produced using the Expected Moments Algorithm as required by USGS Bulletin 17-C.  The results of the modeling are presented in conformance with Engineering Regulation 1105-2-101 and Engineering and Construction Bulletin 2019-11.

Identification of Formulas Used in the Software and Proof that the Computations Are Done Correctly

HEC-FDA uses Monte Carlo sampling techniques to capture natural variability and epistemic uncertainty in the damage with a given probability of exceedance. Most input modeling is defined using a two-dimensional relationship; for example, flow and probability, stage and flow, and stage and damage. Thus, these input relationships have x and y values, where the x is the independent variable (which may be a probability of exceedance, a flow, or a stage) and y is the dependent variable (which may be a distribution of flows, a distribution of stages, or a distribution of dollar damages). The user has the choice of triangular, Normal, Log Normal, and truncated distributions to represent the uncertainty about a variable. The Monte Carlo sampling technique is the defining formula of the HEC-FDA software program. The appropriateness of this approach has been accepted and used by professionals in the Dam and Levee Safety Community of Practice in the Risk Management Center, and in the Modeling and Mapping Center of Expertise. More information on the equations used and the equations themselves are available on GitHub.

System Quality

Description and Rationale for Selection of Supporting Software Tool/Programming Language and Hardware Platform

HEC-FDA is written using C#, XAML, and XML, and compiled and built using the .NET framework and the .NET Core platform. The primary language of the computational engine is C#. C# has been selected because the language is an object-oriented high-level and robust programming language that abstracts away from routine procedures.

Proof that the Programming Was Done Correctly

The primary proof that the programing was done correctly is based on the several hundred unit tests that confirm that each function with a given set of inputs results in the expected output. These tests are available for inspection on GitHub. What's more, every change to the codebase is first tested against every unit test before making the change official, thereby ensuring that enhancements to the codebase do not break existing code. HEC-FDA also underwent HEC's internal review process prior to release. The process consists of a formal review of both the software and user documentation. The process also includes evaluating the results of existing studies. Ten case studies were evaluated by HEC and the FRM-PCX. Review the Version 2.0 Release Notes for more information on the case studies. Any comments or concerns are addressed prior to release. 

Availability of Software and Hardware Required by the Software Program

The Windows and hydraulic engineering software programs compatible with HEC-FDA were chosen based on their widespread application and availability. The HEC-FDA software is available for download here. User documentation is available at HEC-FDA User Documentation and https://github.com/HydrologicEngineeringCenter.

Description of Process Used to Test and Validate Software

HEC-FDA underwent HEC's internal review process prior to release. The process consists of a formal review of both the software and user documentation. The process also includes evaluating the results of existing studies. Ten case studies were evaluated by HEC and the FRM-PCX. Review the Version 2.0 Release Notes for more information on the case studies. Any comments or concerns are addressed prior to release. 

Discussion of the Ability to Import Data Into Other Software Analysis Tools (Interoperability Issue)

HEC-FDA stores data in standard formats, such as .hdf, .shp, and .xml. These file types can be read by many different software across multiple operating systems. Tabular data is easily exportable to Microsoft Excel, or other spreadsheet analysis software. Data available in .shp format can be used with most GIS software programs.

Usability

Availability of the Input Data Necessary to Support Modeling

One of the strengths of HEC-FDA as a software application is the wide availability of input data required for consequence assessment.  The list below summarizes the data input required by HEC-FDA and potential sources.

1. Hydrologic modeling is highly recommended. Hydrologic modeling consists of a flow-exceedance probability function and potentially an inflow-outflow function. This modeling can be accomplished using HEC-HMS and HEC-SSP.
2. Hydraulic modeling is required. Hydraulic modeling consists of a rating curve, a set of 8 hydraulic events, and potentially an exterior-interior relationship. As an alternative to the items in (1), a graphical stage-exceedance probability function can be used. This modeling can be accomplished using HEC-RAS.
3. If a levee is included in the modeling, the top elevation of the levee is required. A geotechnical fragility function is highly recommended. The software that geotechnical engineers can use to accomplish this modeling includes Slope/W, Seep/W, Excel, @Risk, UTexas, and Seep3d.
4. Economic data is required, including a geo-referenced structure inventory with information about the structures' types, values, first floor elevations, and location. Additionally, information about each occupancy type is required, including the uncertainty around the first floor elevation, the structure value, and the content value, and a depth-percent damage function with uncertainty. Structure inventory data is available using the National Structure Inventory or official data sets such as a tax assessor database. Depth-percent damage functions are found in Economic Guidance Memorandum 01-03, 04-01, and 09-04, among other publications.

Formatting of Output in an Understandable Manner

Results can be viewed as histograms and tables. Histograms illustrate the shape of the distribution of the various results, and tabular summaries highlight statistics of interest, such as the mean, first quartile, median, and third quartile. See Scenario Results for an example of the histograms and tables used in report display. 

Usefulness of Results to Support Study Analysis

The results are presented at a level of detail that supports the six-step planning process, primarily the identification of the national economic development plan for flood risk management feasibility, CAP, and other studies. What's more, the results are presented with uncertainty for compliance with Engineering Regulation 1105-2-100 and Engineering Regulation 1105-2-101. The characterization of uncertainty is critical for decision-makers to be able to make decisions confidently. Finally, results will also be presented to assist practitioners in the evaluation of NFIP levee accreditation in compliance with Engineering and Construction Bulletin 2019-2.

Ability to Export Results Into Study Reports

HEC-FDA's use of common, standard data formats makes results easily exportable to Excel or other programs for further analysis and generating tables and figures for reports. Copy and paste works.

Training Availability

USACE, together with CEIWR-HEC, provide training on flood risk assessment with the use of HEC-FDA yearly. This training increases the user's knowledge, proficiency, ability, and skill in the use of the software. Training is also provided to districts and elsewhere as requested. Please see the HEC-FDA Training space for training resources.

User Documentation Availability and Whether It Is User Friendly and Complete

The user documentation is available at HEC-FDA User Documentation. The user documentation has been written in compliance with the Plain Writing Act, Section 508 of the Rehabilitation Act of 1973, as amended (29 U.S.C. 794d), and HEC's standards for formatting and user documentation.

Technical Support Availability

The Hydrologic Engineering Center (CEIWR-HEC) provides technical support for USACE users of HEC-FDA. The customer support email address is monitored (HEC.FDA@usace.army.mil) as well as the HEC-FDA Discourse page, which is an online customer support forum for USACE users of all HEC products (https://discourse.hecdev.net/c/fda/33). Non-USACE users should use an internet search engine to locate a vendor that can provide technical support for HEC software. HEC will respond to instances of documented program errors in HEC software.

Software/Hardware Platform Availability to All or Most Users

The latest version of HEC-FDA is available for download by the public here. USACE employees may download the software from the App Portal. HEC-FDA runs on the widely-used and commercially-available Microsoft Windows operating system and is compatible with 32-bit and 64-bit machines.

Accessibility of the Software

The availability of HEC-FDA, its supporting software and hardware, and required input data make HEC-FDA a highly accessible software program to USACE engineers, economists and the planning community at large.

Transparency of Software and How It Allows for Easy Verification of Calculations and Output

Most importantly, the software code and unit tests are available for inspection to the public at https://github.com/HydrologicEngineeringCenter, the defining element of transparency. The ability to view the results at the impact-area-damage-category-project-condition-analysis-year level with summary statistics allows the user to troubleshoot and/or verify whether the results are reasonable given the input data used.