SEMIP compares a variety of models with each other and with observations on a number of model output variables. Due to the large number of different modeling steps involved, these comparisons are best thought of as occurring on a set of OUTPUT LEVELS. At each level, SEMIP collects model output runs for the various test cases, both from models run in house and from model output that is submitted.
SEMIP collects both model output and observational data at each output level. SEMIP also collects all the information needed to run the test cases.
Analysis output, test case input datasets, and other collected sharable datasets will be accessed through the SEMIP Viewer.
Submission of data to SEMIP is through our data warehouse system that also collects and organizes metadata that can help make the data more accessible, useful, and searchable. Data submitted to SEMIP are governed by our Fair Use Data Policy.
Output Levels and Data
SEMIP stores information from systems and model runs and submitted model output at the following levels. Note that submitted model information may only be available in aggregate or analyzed form depending on the restrictions placed on it by the submitting organization. All information is first stored in the native model format/variables/units, and then converted into standard formats, variables, and units for comparison purposes. Due to system requirements access is only provided to the standard formated data.
Output levels are:
Each of these is defined below.
BASIC FIRE INFORMATION
Although most often considered as input datasets, basic fire information is critically important to correctly model fire consumption, emissions, and smoke. For this reason, SEMIP also cross compares the information available from a number of systems, and stores these comparisons. Additionally, the initial test cases all have defined fire information inputs for use by any models being compared in order to (as much as possible) remove fire information as a variable.
Standard SEMIP basic fire information comparison variables:
- FIRE SIZE (BY DAY)
- FIRE GROWTH (BY DAY)
- LATITUDE/LONGITUDE (CENTROID EACH BURNING AREA, BY DAY, as available)
- FIRE PERIMETER(S) (GIS SHAPEFILE, BY DAY, as available)
- FIRE SEVERITY (GIS SHAPEFILE, as available)
FUEL LOADINGS
Fuel loading maps are needed to obtain fuel loadings when they are not available otherwise. Fuel loadings are typically split by size or other classifications, making the most cross cutting analysis TOTAL FUEL LOADING, and where possible TOTAL LIVE and TOTAL DEAD FUELS. Cross comparisons of CANOPY FUELS, DOWN WOODY BY SIZE, and DUFF are also possible with many of the fuel loading maps.
Standard SEMIP fuel loading comparison variables:
- TOTAL FUELS (BY LOCATION)
- TOTAL LIVE (BY LOCATION, as available)
- TOTAL DEAD (BY LOCATION, as available)
- TOTAL CANOPY (BY LOCATION, as available)
- TOTAL DOWN, WOODY BY SIZE [1/10/100/1000 HR] (BY LOCATION, as available)
- TOTAL DUFF (BY LOCATION, as available)
TOTAL CONSUMPTION
Depending on the model used, consumption may be broken down by phase: flaming, smoldering, or residual. Many models produce rates of consumption over time; these are summed to total consumption for this comparison level.
Standard SEMIP total comparison variables:
- TOTAL CONSUMPTION
- TOTAL FLAMING CONSUMPTION (as available)
- TOTAL SMOLDERING CONSUMPTION (as available)
- TOTAL RESIDUAL CONSUMPTION (as available)
- FUEL MOISTURES BY SIZE CLASS (as available)
TIME PROFILE OF CONSUMPTION
Often a companion step to TOTAL CONSUMPTION, time rate of consumption is split out as a separate level because it adds a new dimension (time), and a different class of models can provide this information (e.g. simple idealized time profiles). To cross compare, all of the time rates are normalized against the total amount of consumption, creating a simple fraction of the total amount consumed or total amount consumed by the fire phase in question.
Standard SEMIP time rate of consumption variables:
- FRACTION OF TOTAL CONSUMPTION BY HOUR
- FRACTION OF TOTAL FLAMING CONSUMPTION BY HOUR (as available)
- FRACTION OF TOTAL SMOLDERING CONSUMPTION BY HOUR (as available)
- FRACTION OF TOTAL RESIDUAL CONSUMPTION BY HOUR (as available)
SPECIATED EMISSIONS
Once time rates of consumption are known, speciated emissions, including heat released, are calculated. While some models contain advanced dynamics which change the emissions factors based on, e.g., the amount of heat released, most perform simple emissions factor multiplications, sometimes differentiated by fire phase.
Standard SEMIP speciated emissions comparison variables:
- TOTAL RELEASED PER AMOUNT CONSUMED BY SPECIES
-
HEAT, TOTAL CARBON, CO2, NOx, GHG EQUIVALENTS, CH4, SOx, other species as available
VERTICAL PLUME PROFILES
Vertical plume profiles are created by plume rise models that range from the simple to very complex calculations involving atmospheric stability.
Standard SEMIP vertical comparison variables:
- FRACTION OF TOTAL EMISSIONS BY HEIGHT
- FRACTION OF TOTAL EMISSIONS BY SPECIES BY HEIGHT (as available)
TOTAL COLUMN SMOKE
After dispersion of the plume by dispersion models, total integrated columar smoke measures the overall horizontal advection and diffusion, as well as deposition effects.
Standard SEMIP total column smoke variables:
- TOTAL INTEGRATED SMOKE [in quantity per m2]
- TOTAL INTEGRATED SMOKE BY LAYER [MIXING HEIGHT, TROPOPAUSE]
- AEROSOL OPTICAL DEPTH
GROUND CONCENTRATIONS
The most important values of a smoke dispersion system are the predicted concentrations. These can reveal a substantially different pattern than TOTAL COLUMN SMOKE due to positioning of the plume within the vertical column and local terrain effects, such as slope drainage flows or small valley to large-scale inversions.
Standard SEMIP ground concentration variables:
- PARTICULATE MATTER [PM2.5, PM10, PM TOTAL] SURFACE CONCENTRATIONS
- OTHER SPECIES (as available) [OZONE, NOx, SOx, PH] SURFACE CONCENTRATIONS
- SURFACE VISIBILITY (as available)
Observational Datasets
A large collection of observational datasets have been identified
and are being collected by SEMIP. These serve not only to provide
"ground-truth" information on which to evaluate model performance, but
also can, in some cases, serve as inputs to additional modeling steps.
See also:
Test Case Inputs
Because of the need to be able to both run the test cases on various models internally and enable outside participants to run the test cases remotely and submit their results, SEMIP collects and stores all the needed test case input data. These include fire information, atmospheric model data (for fuel moisture calculations and to run dispersion models), and various intermediate level datasets needed to provide a baseline for the next model level. The later includes fuel loading information, total consumption, total consumption by hour, and vertical plume profiles. More information can be found in the Test Cases section.
SEMIP Viewer
The SEMIP Viewer is under construction. This web tool will provide users with interactive access to SEMIP datasets and analyses.
Data
