Post-fire flood hazard model (PF2HazMo) version 1.0.0: Model scripts and parameterization and validation data
Data files
Abstract
Human development at the foot of the mountains faces sediment-laden flood hazards characterized by high-velocity, erosive flows carrying mud and debris, and when flood control infrastructure that protects communities fills with sediment, it loses capacity. The estimation and management of sediment-laden floods have proven challenging because cycles of wildfire, precipitation, and infrastructure sedimentation are still poorly understood. Efforts to model compound hazards such as post-fire floods are relatively new, and existing models do not consider the role of flood control infrastructure, such as debris retention basins and flood channels, in the development of post-fire floods. Here we present data sources and calibration methods to estimate sediment-laden flood hazards downstream of infrastructure on a catchment-by-catchment basis using the Post-Fire Flood Hazard Model (PF2HazMo), a stochastic modeling approach that utilizes continuous simulation to resolve the effects of antecedent conditions and system memory. Data sources provide parameter ranges needed for stochastic modeling, and several performance measures are considered for model calibration. With application to three catchments in Southern California, we show that PF2HazMo predicts the median of the simulated distribution of peak bulked flows within the 95% confidence interval of observed flows, with an order of magnitude range in bulked flow estimates depending on the performance measure used for calibration. Using infrastructure overtopping data from a post-fire wet season, we show that PF2HazMo accurately predicts the number of flood channel exceedances. Model applications to individual watersheds reveal whether existing infrastructure is undersized to contain present-day and future overtopping hazards based on current design standards.
README: Post-Fire Flood Hazard Model (PF2HazMo) Version 1.0.0: Model Scripts and Parameterization and Validation Data
https://doi.org/10.5061/dryad.8w9ghx3t2
This repository contains the model scripts and data used to parameterize, calibrate, and validate the Post-Fire Flood Hazard Model (PF2HazMo). PF2HazMo is a stochastic modeling framework that utilizes continuous simulation to resolve interactions between cycles of wildfire, rainfall, and the filling of protective flood control infrastructure with sediment. Parameterization refers to estimating values for model parameters using data, while calibration refers to adjusting the value of a given model parameter to produce the closest fit between simulations and observations of the model response variable--in this case, post-fire peak streamflow. By "validation," we mean the evaluation of model performance by comparing simulation results to observations of the model response variable--in this case, flood control infrastructure exceedances.
This repository also contains model scripts and data used to conduct a sensitivity analysis and error propagation analysis of the model. Sensitivity analysis was used to identify the model parameters for which a small change in their value would produce a relatively large change in key model response variables. Error propagation analysis was conducted to determine the relationship between changes in model parameters and changes in key model response variables taking into account the combined influence of all parameters simultaneously. Both analyses are useful for guiding model parameter estimation and calibration strategies and for developing a better understanding of how model parameters, inputs, and outputs affect one another. For the publication associated with this repository, the sensitivity analysis and error propagation analysis were used to identify which model parameters to calibrate, represent as random variables, or parameterize.
Finally, the repository contains model scripts used to generate post-fire flood hazard projections after model parameterization and calibration. These scripts may be used to apply the modeling framework to a specific watershed and produce a long-term forecast of bulked flows (mixture of water and sediment volumetric flows) and infrastructure exceedances under user-specified wildfire frequency/severity and infrastructure design/maintenance scenarios. All scripts produce time series at either daily or annual intervals.
The model was developed using and is designed to run in MATLAB. A detailed description of the methods used to retrieve data and parameterize, calibrate, and validate PF2HazMo is provided in the publication associated with this repository.
Description of the data and file structure
List of Contents:
- Folder: Calibration_and_Validation_Codes (contains subfolder "Recovery_Timescale_Estimation")
- Folder: Data
- Folder: Projection_Codes
- Folder: Sensitivity_and_Error_Analysis_Codes
- File: CGS_Holy Fire Discharge Estimates_11-29-2018 and 12-06-2018 storms_4-22-2024_FINAL.pdf
Detailed description of folder contents
1. Calibration_and_Validation_Codes:
File List:
- calibfreq_nce_nbe_svbend_sfb_sfc_veb_ffb_ffc_McVicker_2018_19.mat
- Coldwater Canyon Rainfall Data 340 11012018-03312019_formatted.xlsx
- Coldwater_Nov2018_Mar2019_DailyPrecip_mm.mat
- DayToMonth.mat
- export_results_to_CSV.m
- Leach_Nov2018_Mar2019_DailyPrecip_mm.mat
- Maximum_Daily_Runoff_cms_USGS_11071760_20181204_20190331.csv
- McVicker_Nov2018_Mar2019_DailyPrecip_mm.mat
- modelparams_v2.mat
- ParamPrep.m
- PF2HazMo_Calib_Coldwater.m
- PF2HazMo_Param_Leach.m
- PF2HazMo_Param_McVicker.m
- returnlvls_v2.mat
- simdates.mat
Relationship between files, if important:
- ParamPrep.m requires the following codes and/or data files to run (i.e., these are dependencies of the script): Coldwater Canyon Rainfall Data 340 11012018-03312019_formatted.xlsx
- Note: this data file is included as an example observed precipitation time series to help the user better understand how the script formats precipitation data for use as model input
- PF2HazMo_Calib_Coldwater.m requires the following dependencies: Coldwater_Nov2018_Mar2019_DailyPrecip_mm.mat, DayToMonth.mat, Maximum_Daily_Runoff_cms_USGS_11071760_20181204_20190331.csv, modelparams_v2.mat, returnlvls_v2.mat, simdates.mat
- PF2HazMo_Param_Leach.m requires the following dependencies: Leach_Nov2018_Mar2019_DailyPrecip_mm.mat, DayToMonth.mat, modelparams_v2.mat, returnlvls_v2.mat, simdates.mat
- PF2HazMo_Param_McVicker.m requires the following dependencies: McVicker_Nov2018_Mar2019_DailyPrecip_mm.mat, DayToMonth.mat, modelparams_v2.mat, returnlvls_v2.mat, simdates.mat
- export_results_to_CSV.m requires the following dependencies: calibfreq_nce_nbe_svbend_sfb_sfc_veb_ffb_ffc_McVicker_2018_19.mat
- Note: this data file is included as an example model output to help the user better understand how the script formats the model results
Description of folder contents:
- Contains MATLAB scripts and data files needed to calibrate the model using USGS streamgage data from Coldwater Watershed ("PF2HazMo_Calib_Coldwater.m"); i.e., this script is used to calculate the optimal value of the Runoff Response parameter using a set of four objective functions: Nash-Sutcliffe Efficiency (NSE), Root Mean Squared Error (RMSE), Percent Error (pct_err), and Mean Absolute Error (MAE).
- Contains MATLAB scripts and data files needed to run the calibrated model for Leach Watershed ("PF2HazMo_Param_Leach.m") and McVicker Watershed ("PF2HazMo_Param_McVicker.m") for the first wet season following the Holy Fire. Simulation results from these models were compared to observed infrastructure exceedances to evaluate model performance.
- The script "ParamPrep.m" is designed to format rain gauge data (found in "Data" folder) such that it may be used as input/forcing data for simulations.
- The script "export_results_to_CSV.m" is designed to facilitate the export of simulation results to a tabular format (in this case, a Comma Separated Values file).
Subfolder: "Recovery_Timescale_Estimation":
File List:
- Coldwater_Watershed_11071760_Fire_History.xlsx
- Coldwater2000_2023_Mean_EVI_NDVI.csv
- estimate_recovery_timescale2_Coldwater.m
- estimate_recovery_timescale2_LeachMcVicker.m
- FRAP_allPerims_2010_2021_RivCo_NoDups.csv
- LeachMcVicker2000_2023_Mean_EVI_NDVI.csv
- mannkendall.m
Relationship between files, if important:
- estimate_recovery_timescale2_Coldwater.m requires the following codes and/or data files to run (i.e., these are dependencies of the script): Coldwater_Watershed_11071760_Fire_History.xlsx, Coldwater2000_2023_Mean_EVI_NDVI.csv, mannkendall.m
- estimate_recovery_timescale2_LeachMcVicker.m requires the following dependencies: FRAP_allPerims_2010_2021_RivCo_NoDups.csv, LeachMcVicker2000_2023_Mean_EVI_NDVI.csv, mannkendall.m
Description of folder contents:
- Contains MATLAB scripts and data files needed to estimate the recovery timescale parameter for Coldwater Watershed ("estimate_recovery_timescale2_Coldwater.m") and that for Leach and McVicker Watersheds ("estimate_recovery_timescale2_LeachMcVicker.m").
2. Data:
File List:
- Coldwater Canyon Rainfall Data 340 11012018-03312019_formatted.xlsx
- Coldwater_Parameter_Values.xlsx
- Leach - Dickey Canyon 344_formatted.xlsx
- Leach and McVicker Sediment Excavation Records_m3.csv
- Leach_Parameter_Values.xlsx
- Maximum_Daily_Runoff_cms_USGS_11071760_20181204_20190331.csv
- McVicker Canyon 343_formatted.xlsx
- McVicker_Parameter_Values.xlsx
- USGS 11071760 COLDWATER_instantaneous discharge cfs_20181204_20230227.xlsx
Relationship between files, if important: None
Description of folder contents:
- Rain gauge data is stored in spreadsheets "Coldwater Canyon Rainfall Data 340 11012018-03312019_formatted.xlsx", "Leach - Dickey Canyon 344_formatted.xlsx", and "McVicker Canyon 343_formatted.xlsx." These data were provided by Riverside County Flood Control and Water Conservation District with written permission to share with the public.
- Runoff data for USGS streamgage 11071760 is stored in two files: the raw data at 15-minute intervals in units of cfs are stored in "USGS 11071760 COLDWATER_instantaneous discharge cfs_20181204_20230227.xlsx" while the maximum daily runoff values in units of cms are saved in "Maximum_Daily_Runoff_cms_USGS_11071760_20181204_20190331.csv"
- Model parameter values and intermediate calculations for parameter estimation are stored in three spreadsheets, one for each watershed: "Coldwater_Parameter_Values.xlsx", "Leach_Parameter_Values.xlsx", and "McVicker_Parameter_Values.xlsx"
- Excavation (i.e., sediment removal) data for the Leach and McVicker debris basins are stored in spreadsheet "Leach and McVicker Sediment Excavation Records_m3.csv." These data were provided by Riverside County Flood Control and Water Conservation District with written permission to share with the public.
3. Projection_Codes:
File List:
- DayToMonth.mat
- export_results_to_CSV.m
- genprcp.m
- MCMCv2_100yrsrlzn_EVA.mat
- output_Leach.mat
- output_Leach.txt
- PF2HazMo_Forecast_Leach.m
- PF2HazMo_Forecast_McVicker.m
- pfireocc.m
Relationship between files, if important:
- PF2HazMo_Forecast_Leach.m requires the following codes and/or data files to run (i.e., these are dependencies of the script): DayToMonth.mat, genprcp.m, MCMCv2_100yrsrlzn_EVA.mat, pfireocc.m
- PF2HazMo_Forecast_McVicker.m requires the dependencies: DayToMonth.mat, genprcp.m, MCMCv2_100yrsrlzn_EVA.mat, pfireocc.m
- export_results_to_CSV.m requires the dependencies: output_Leach.mat
- Note: this data file is included as an example model output to help the user better understand how the script formats the model results
Description of folder contents:
- Contains MATLAB scripts and data files needed to run the calibrated model in "forecast mode" for Leach Watershed ("PF2HazMo_Forecast_Leach.m") and McVicker Watershed ("PF2HazMo_Forecast_McVicker.m") for 100 years considering two (2) scenarios: a fire interval of 40 years (representing the present fire regime) and a fire interval of 20 years (representing a potential future fire regime).
- The file "MCMCv2_100yrsrlzn_EVA.mat" is a synthetic precipitation time series (100 years, daily interval) produced using the Monte Carlo Markov Chain (MCMC) stochastic rainfall generator described in Jong-Levinger et al. (2022).
- The script "export_results_to_CSV.m" is designed to facilitate the export of simulation results to a tabular format (in this case, a Comma Separated Values file).
- The file "output_Leach.mat" is an example model output from "PF2HazMo_Forecast_Leach.m" that is used as an input for "export_results_to_CSV.m". "output_Leach.txt" is an example of the model configuration file that can be generated when the model is run in forecast mode by setting "options_exprtconfig" equal to 1; it is a text file containing user-specified options and parameter values for a given simulation trial.
4. Sensitivity_and_Error_Analysis_Codes:
File List:
- DayToMonth.mat
- ea_param_vals.mat
- Leach - Dickey Canyon 344_formatted.xlsx
- Leach and McVicker Sediment Excavation Records_m3.csv
- McVicker Canyon 343_formatted.xlsx
- OneYear_DailyPrecip_mm_Jun1start.mat
- ParamPrep_EA.m
- ParamPrep_SA.m
- PF2HazMo_ErrorAnalysis.m
- PF2HazMo_SensitivityAnalysis.m
- returnlvls_v2.mat
- sa_param_vals.mat
- simdates.mat
- storm_statistics_from_gauge_record_byWetSeason.m
- stormdur_hr_lch_mv_2018-21_all.mat
Relationship between files, if important:
- ParamPrep_EA.m requires the following codes and/or data files to run (i.e., these are dependencies of the script): Leach and McVicker Sediment Excavation Records_m3.csv, stormdur_hr_lch_mv_2018-21_all.mat, returnlvls_v2.mat
- ParamPrep_SA.m requires the dependencies: stormdur_hr_lch_mv_2018-21_all.mat, returnlvls_v2.mat
- PF2HazMo_ErrorAnalysis.m requires the dependencies: DayToMonth.mat, ea_param_vals.mat, OneYear_DailyPrecip_mm_Jun1start.mat, simdates.mat
- PF2HazMo_SensitivityAnalysis.m requires the dependencies: DayToMonth.mat, sa_param_vals.mat, OneYear_DailyPrecip_mm_Jun1start.mat, simdates.mat
- storm_statistics_from_gauge_record_byWetSeason.m requires the dependencies: Leach - Dickey Canyon 344_formatted.xlsx, McVicker Canyon 343_formatted.xlsx
- Note: this script can be used to output a dataset of storm durations from rain gauge data (example data file included is "stormdur_hr_lch_mv_2018-21_all.mat") that can be used to estimate the time to peak parameter (tp) for the error propagation analysis using the script "ParamPrep_EA.m"
Description of folder contents:
- Contains MATLAB scripts and data files needed to conduct the sensitivity analysis ("PF2HazMo_SensitivityAnalysis.m") and the error propagation analysis ("PF2HazMo_ErrorAnalysis.m").
- The script "storm_statistics_from_gauge_record_byWetSeason.m" contains a storm identification algorithm used to calculate storm duration from rain gauge records. The resulting empirical distribution of storm durations can be used to estimate the time to peak parameter (tp) for the error propagation analysis using the script "ParamPrep_EA.m". "storm_statistics_from_gauge_record_byWetSeason.m" can also output total precipitation depth and precipitation intensity per storm.
Variable list for model scripts
This section contains variable lists pertaining to each model script in the repository. Each variable list in this section defines variables not covered by the preceding variable list (to prevent repetition of variables used across all scripts).
PF2HazMo_Calib_Coldwater.m, PF2HazMo_Param_Leach.m, PF2HazMo_Param_McVicker.m
Variable List:
- alpha: fraction by which debris basin attenuates flood peak, with a value of 1 representing no flood peak attenuation (unitless)
- B: bulked flows (combination of water and sediment flows) from watershed outlet (m3/s)
- Crational: dimensionless runoff coefficient from Rational Method; represents ratio of precipitation to area-normalized runoff (unitless)
- c_tp: ratio of dimensionless runoff coefficient from Rational Method to time to peak (1/seconds)
- clean_rate: daily excavation rate for debris basin during wet season (m3/day)
- clean_thresh: fraction of flood control infrastructure volume filled with sediment that triggers cleaning (unitless)
- daysinyr: number of days in one year of the simulation; if simulation lasts less than a year, set "daysinyr" to less than 365 (days)
- firestartdate: number of days into simulation when fire should occur (days)
- J_c: sediment flux into flood channel (m3)
- k: bulking factor; ratio of sum of water and sediment flows to water flow alone (unitless)
- k_design: design bulking factor used to size flood control infrastructure (unitless)
- k0: baseline bulking factor when watershed is unaffected by fire (unitless)
- k1: post-fire bulking factor; bulking factor value immediately after fire occurs (unitless)
- k1rng: two-element array of post-fire bulking factors representing range of possible bulking factors produced by a given wildfire; when wildfire occurs, post-fire bulking factor is uniformly randomly sampled from this range (unitless)
- MAE: Mean Absolute Error (m3/s)
- NSE: Nash-Sutcliffe Efficiency (unitless)
- nsims: number of simulations per run
- nyears: number of years in simulation; if simulation lasts less than a year, set to 1 (years)
- obsQ: array of observed peak flows from watershed (includes bulked flows) (m3/s)
- P: precipitation time series (daily interval); used as model forcing/input (mm)
- P_channel_design: magnitude of design storm used to size flood control infrastructure (mm)
- pct_err: percentage error (%)
- Q_c: volumetric discharge into flood channel; includes both clear-water and bulked flows (m3/s)
- Q_c_max: effective channel capacity (as a volumetric discharge) (m3/s)
- Q_c_max_clean: capacity of clean flood channel (as a volumetric discharge) (m3/s)
- Q_hazard: magnitude of overbank flows (i.e., flows in excess of flood channel capacity); includes both clear-water and bulked flows (m3/s)
- Q_w: clear-water (or "water only") volumetric discharge from watershed outlet (m3/s)
- RMSE: Root Mean Squared Error (m3/s)
- simdates: array of simulation dates (this variable is only needed when simulation duration is less than one year)
- simpkQ: array of simulated peak flows from watershed on dates corresponding to those of observed peak flows (includes bulked flows) (m3/s)
- t: time (days)
- T_recovery: recovery timescale; time scale for watershed vegetation recovery after fire (days)
- tp: time to peak; duration from beginning of storm to time of peak runoff (seconds)
- Vol_b: volume of sediment in debris basin (m3)
- Vol_b_max: capacity of clean debris basin (m3)
- Vol_c: volume of sediment in flood channel (m3)
- Vol_c_max: capacity of clean flood channel (as a volume) (m3)
- Vol_s: volume of sediment from watershed outlet (m3)
- Vol_s_remov_basin: volume of sediment removed from debris basin due to excavation (m3)
- Vol_s_remov_chan: volume of sediment removed from flood channel due to excavation (m3)
- Vol_w: volume of water from watershed outlet (m3)
- w: wet season cleaning waiting period; number of days with no rain needed before debris basin cleaning can commence (days)
- WatershedArea: contributing drainage area (km2)
PF2HazMo_Forecast_Leach.m, PF2HazMo_Forecast_McVicker.m
Variable List (additional variables not covered above):
- p_fire: daily probability that fire will occur during fire season (unitless)
- RIfire: array of fire return intervals that will be used in each simulation scenario (years)
PF2HazMo_ErrorAnalysis.m, PF2HazMo_SensitivityAnalysis.m
Variable List (additional variables not covered above):
- nparams: number of model parameters included in analysis
- nvals: number of parmeter values to test in simulation
- noutputs: number of output values to be included in test results (this must be user-specified)
- nce_mat: array of number of flood channel exceedances
- nbe_mat: array of number of debris basin exceedances
- tobf_mat: array of sum of volume of bulked flows exceeding channel capacity (m3)
- maxobf_mat: array of maximum volume of bulked flows exceeding channel capacity (m3)
- tfc_mat: array of sum of flows into flood channel; includes both clear-water and bulked flows (m3)
- tfb_mat: array of sum of flows into debris basin; includes both clear-water and bulked flows (m3)
Data-specific information by dataset
Coldwater Canyon Rainfall Data 340 11012018-03312019_formatted.xlsx
- Number of variables: 3
- Number of cases/rows: 1084
- Variable List:
- Gauge_ID: rain gauge ID number
- Datetime: date and time of measurement (mm/dd/yyyy HH:mm)
- Precip_in: precipitation depth (inches)
- Missing data codes: None
- Specialized formats or other abbreviations used: None
Coldwater_Parameter_Values.xlsx, Leach_Parameter_Values.xlsx, McVicker_Parameter_Values.xlsx
- Variable List:
- Watershed area: contributing drainage area (square kilometers)
- Runoff coefficient / Time to peak (c/tp): ratio of dimensionless runoff coefficient from Rational Method to time to peak (1/seconds)
- Time to peak (tp): duration from beginning of storm to time of peak runoff (seconds)
- Recovery timescale (tr): time scale for watershed vegetation recovery after fire (days)
- Cleaning threshold: fraction of flood control infrastructure volume filled with sediment that triggers cleaning (unitless)
- Debris basin flood peak attenuation coefficient (α): fraction by which debris basin attenuates flood peak, with a value of 1 representing no flood peak attenuation (unitless)
- Wet season cleaning waiting period (w): number of days with no rain needed before debris basin cleaning can commence (days)
- Wet season cleaning rate: daily excavation rate for debris basin during wet season (m3/day)
- Debris basin design capacity: storage capacity of debris basin when clean (m3)
- Flood channel design peak flow rate: flow capacity of flood channel when clean (m3/s)
- Post-fire bulking factor (k) range: range of values from which to sample stochastically for the post-fire bulking factor, the ratio of the combined water and sediment volumetric flux to the water volumetric flux (unitless)
- Missing data codes: NA (not applicable), #N/A (data not available)
- Specialized formats or other abbreviations used: EVI = Enhanced Vegetation Index, NDVI = Normalized Difference Vegetation Index, HEC-HMS = Hydrologic Engineering Center Hydrologic Modeling System, CN = Curve Number, Vw = water volumetric flux, Vs = sediment volumetric flux, NSE = Nash-Sutcliffe Efficiency, RMSE = Root Mean Squared Error, MAE = Mean Absolute Error
Leach - Dickey Canyon 344_formatted.xlsx
- Number of variables: 3
- Number of cases/rows: 3330
- Variable List:
- GaugeID: rain gauge ID number
- DateTime: date and time of measurement (mm/dd/yyyy HH:mm:ss)
- Precip_in: precipitation depth (inches)
- Missing data codes: None
- Specialized formats or other abbreviations used: None
Leach and McVicker Sediment Excavation Records_m3.csv
- Number of variables: 3
- Number of cases/rows: 67
- Variable List:
- Basin: Debris basin name
- Date: date of sediment excavation (mm/dd/yyyy)
- Sediment_m3: volume of sediment removed (cubic meters)
- Missing data codes: None
- Specialized formats or other abbreviations used: None
Maximum_Daily_Runoff_cms_USGS_11071760_20181204_20190331.csv
- Number of variables: 2
- Number of cases/rows: 119
- Variable List:
- datetime: date (mm/dd/yyyy)
- Q_cms: maximum daily runoff (m3/s)
- Missing data codes: None
- Specialized formats or other abbreviations used: None
McVicker Canyon 343_formatted.xlsx
- Number of variables: 3
- Number of cases/rows: 2918
- Variable List:
- GaugeID: rain gauge ID number
- DateTime: date and time of measurement (mm/dd/yyyy HH:mm:ss)
- Precip_in: precipitation depth (inches)
- Missing data codes: None
- Specialized formats or other abbreviations used: None
USGS 11071760 COLDWATER_instantaneous discharge cfs_20181204_20230227.xlsx
- Number of variables: 6
- Number of cases/rows: 146414
- Variable List:
- agency_cd: agency code
- site_no: site number
- datetime: date and time of measurement (mm/dd/yyyy HH:mm:ss)
- tz_cd: time zone code
- Q_cfs: discharge (ft3/s)
- QAcode: Quality assurance code - abbreviation definitions explained under "Specialized formats or other abbreviations used"
- Missing data codes: None
- Specialized formats or other abbreviations used: PST = Pacific Standard Time, A = Approved for publication -- Processing and review completed, P = Provisional data subject to revision, e = Value has been estimated
Sharing/Access information
- Licenses/restrictions placed on the data: CC0
- Licenses/restrictions placed on the software: GNU General Public License v3.0 or later
- Licenses/restrictions placed on the supplemental information: Creative Commons Attribution 4.0 International (CC BY 4.0) license
- Publications that cite or use the data: Jong-Levinger, A., Houston, D., & Sanders, B. F. (2023). Estimating Post-Fire Flood Infrastructure Clogging and Overtopping Hazards. Water Resources Research.
- Data was derived from the following sources:
- The topographic data used to estimate watershed morphology are publicly available via https://www.usgs.gov/the-national-map-data-delivery.
- The land cover data used to estimate the curve number are publicly available at the Multi-Resolution Land Characteristics (MRLC) consortium via https://www.mrlc.gov/data.
- The soil properties data used to estimate the curve number are publicly available at the Web Soil Survey via https://websoilsurvey.nrcs.usda.gov/app/WebSoilSurvey.aspx.
- The spatially distributed precipitation frequency estimates used to estimate the runoff response parameter and post-fire bulking factors are publicly available at the Precipitation Frequency Data Server via https://hdsc.nws.noaa.gov/pfds/pfds_gis.html.
- The watershed-scale discharge frequency estimates used to estimate the runoff response parameter and post-fire bulking factors are publicly available at the StreamStats website via https://streamstats.usgs.gov/ss/.
- The vegetation indices used to estimate the recovery timescale are publicly available at Google Earth Engine via https://developers.google.com/earth-engine/datasets/catalog/MODIS_061_MOD13Q1.
- The precipitation record used to develop the stochastic rainfall model are publicly available from the Climate Data Online database via https://www.ncei.noaa.gov/cdo-web/search?datasetid=GHCND.
- The rain gauge and debris basin excavation data were provided by Riverside County Flood Control and Water Conservation District with written permission to share with the public.
- The memorandum with discharge estimates was provided by the California Geological Survey with written permission to share with the public as part of this data repository.
Code/Software
MATLAB version 9.13.0.2049777 (R2022b) was used to write and run the model scripts. The Statistics and Machine Learning toolbox must be installed to run some scripts.
Methods
This repository contains the model scripts and data used to parameterize, calibrate, and validate the Post-Fire Flood Hazard Model (PF2HazMo). PF2HazMo is a stochastic modeling framework that utilizes continuous simulation to resolve interactions between cycles of wildfire, rainfall, and the filling of protective flood control infrastructure with sediment. This repository also contains model scripts and data used to conduct a sensitivity analysis and error propagation analysis of the model, as well as to produce post-fire flood hazard projections after model calibration.
The model was developed using and is designed to run in MATLAB.
A detailed description of the methods used to retrieve data and parameterize, calibrate, and validate PF2HazMo is provided in the publication associated with this dataset. Below are links to the publicly available databases from which data was retrieved to parameterize and calibrate the model:
The topographic data used to estimate watershed morphology are publicly available via https://www.usgs.gov/the-national-map-data-delivery. The land cover data used to estimate the curve number are publicly available at the Multi-Resolution Land Characteristics (MRLC) consortium via https://www.mrlc.gov/data. The soil properties data used to estimate the curve number are publicly available at the Web Soil Survey via https://websoilsurvey.nrcs.usda.gov/app/WebSoilSurvey.aspx. The spatially distributed precipitation frequency estimates used to estimate the runoff response parameter and post-fire bulking factors are publicly available at the Precipitation Frequency Data Server via https://hdsc.nws.noaa.gov/pfds/pfds_gis.html. The watershed-scale discharge frequency estimates used to estimate the runoff response parameter and post-fire bulking factors are publicly available at the StreamStats website via https://streamstats.usgs.gov/ss/. The vegetation indices used to estimate the recovery timescale are publicly available at Google Earth Engine via https://developers.google.com/earth-engine/datasets/catalog/MODIS_061_MOD13Q1. The precipitation record used to develop the stochastic rainfall model are publicly available from the Climate Data Online database via https://www.ncei.noaa.gov/cdo-web/search?datasetid=GHCND. The rain gauge and debris basin excavation data were provided by Riverside County Flood Control and Water Conservation District with written permission to share with the public. The data repository also includes supplemental information: a memorandum with discharge estimates from two storms following the 2018 Holy Fire that are referenced in the Discussion section of the publication associated with this data repository; this memorandum was provided by the California Geological Survey with written permission to share with the public as part of this data repository.