1 Introduction

This memo presents ideas for the development of an emissions modeling preprocessor and QA tool. The now defunct SMOKE Tool was a program that provided a series of utilities for preprocessing and viewing different emissions datasets to prepare them for use in SMOKE. We propose resurrecting the concept of a SMOKE preprocessor in the form of a Java program that will expand upon the functions of the original SMOKE Tool by integrating features from UAM-Guides? and EMS-95 and adding new QA and preprocessing utilities. Pulling together a series of current QA utilities and developing a set of new programs for displaying and manipulating emissions data, this Java tool will “close the loop” in the analysis between SMOKE input and output data. The preprocessors in this tool will automate the creation of SMOKE inputs and also make the preparation of SMOKE inputs more accessible to the emissions modeling community.

2 Emissions QA Features

The Emissions QA tool will read raw SMOKE inputs, display the data graphically integrated with different types of GIS information, and automate the production of different QA metrics for evaluating the quality of the data. We propose the features below to be contained in the prototype of this tool.

· Expanding upon an emissions regression prototype program developed for the WRAP, the QA tool will automate the creation of emissions regressions by state, county, and SCC.

· Mapping tool for comparing reported and modeled emissions densities will allow the comparison of SMOKE emissions normalized to land area (tons/sq.meter) to reference information reported in the National Air Pollutant Emissions Trends reports.

· Inventory county maps (similar to EMS-95 plots) by SCC, SCC group, population normalization, pollutant, etc. for comparison of the raw emissions data to post-SMOKE PAVE plots.

· GIS overlays onto PAVE-like emissions plots will produce both emissions density maps and conventional emissions plots overlaid with GIS information

· SMOKE surrogate maps for comparing against GIS shape files. The production of SMOKE surrogate maps will allow the graphical comparison between what SMOKE is using to spatially allocate emissions and the GIS shape files used to create these surrogates.

· Emissions/population normalization and sorting and ranking by a specified metric. This function will automate the creation of cumulative emissions totals for probing the top 80 or 90% of the emissions sources when performing QA.

· Layer fractions QA tool to sort the layer reports coming out of Smkreport and present the information in a more user-friendly format.

· Point source coordinate and stack parameter QA to test the location of stacks relative to county boundaries and to perform sanity checks on stack parameters by SCC class, respectively.

· Tools for summing modeled emissions across different spatial and temporal resolutions

3 Emissions Web Analysis Package

Instrument the QA tool to automate the generation and web-posting of emission QA graphics. Add a utility to the QA tool that would generate a series of graphical and tabular metrics and automatically post the results to a generic web page format. Sample metrics to include on these pages would include Max-0 tile plots for all source categories and pollutants, true Max tile plots for all source categories and pollutants, time series at different spatial and temporal resolutions, inventory regressions, SCC tile plots, surrogate tile plots, temporal profile plots, vertical emissions profile plots, tabular inventory summaries (from Smkreport), documentation files (e.g. spreadsheets, Word docs, etc), and Bugzilla links.

4 Community Analysis Tools

There was a wide-range of interest at the 2003 CMAS Models-3 Workshop in the availability of community analysis tools for emissions modeling. This interest generated discussion about tools developed independently around the SMOKE modeling community for evaluating emissions modeling with SMOKE. Gail Tonessen at UC-Riverside offered a set of tools developed under WRAP funding to the community for auditing SMOKE emissions modeling. These tools post-process SMOKE emissions to create time series, horizontal and vertical spatial aggregations, and temporally averaged emissions reports and graphics independent of SMOKE. We propose to survey the SMOKE modeling community to determine what analysis tools are available that have been independently developed and collect these tools to distribute in an analysis suite for SMOKE evaluation and QA. These utilities could be coupled to or integrated into the emissions QA tool for distribution with SMOKE.

5 Spatial Allocator Enhancements

The Multimedia Integrated Modeling System (MIMS) uses a program called the Spatial Allocator to display modeling grids and to generate spatial surrogates for emissions modeling. We propose modifying the Spatial Allocator to interpolate and window available spatial surrogates and land use information. The Spatial Allocator will be enhanced with a SMOKE surrogate and land use reader to allow these data to be imported into the system for manipulation. Once read in by the Spatial Allocator, windowing of both types of data could be performed to create subdomains from large unified datasets. In addition to windowing the data, the Spatial Allocator could also function to interpolate coarse grid data to a finer grid. Also by reading in land use information, the Spatial Allocator could be used to manipulate and project the land use categories into the future for sensitivities on how emissions are affected by land use changes.

6 SMOKE Biogenic Land Use Preprocessor

SMOKE uses gridded land use files as inputs to the biogenic emissions program BEIS3. As the gridded land use files contain domain-specific information, they must be created for new spatial configurations of SMOKE. Currently these files are prepared in a preprocessing step independent of SMOKE using script-based programs developed by the CEP. To automate the process of creating land use files and to make the preparation of BEIS3 inputs more accessible, we propose adding a BELD à BEIS3 preprocessor to the Spatial Allocator feature of the QA tool. This preprocessor will compile and grid the 1-km BELD land use files into a grid specified by the user. The land use reader to the Spatial Allocator will be written to allow the integration of other land use datasets in the future. (Note: This work has been funded as part of a separate project.)

7 GIS-based Inventory Viewer

Emissions QA requires that the data be reviewed at all key steps in the process of converting raw ASCII inventory files into binary model-ready emissions files. While SMOKE offers a QA package in the program Smkreport that allows monitoring this conversion process, the SMOKE environment lacks a tool for both checking the data before it is used in SMOKE and for auditing SMOKE modeling. A GIS-based inventory viewer that is independent of SMOKE is proposed that will allow modelers to visualize raw inventory files, ancillary input data for cross comparisons between SMOKE input and output data. Adding GIS capabilities to this tool will also allow the layering of different types of information in a graphical display. For example, emissions inventory estimates, a specific land use category assigned by SMOKE, and population density could be plotted together to evaluate the quality of the inventory and spatial surrogate data. Specifically the proposed visualization tool would read IDA and NIF formatted inventory files and display the data graphically by state, county, and/or SCC totals.

8 Projection and Control Packet Interface

SMOKE users need a way to systematically create growth and control packets for projecting emissions and creating inventories for emissions sensitivity studies. The EGAS system creates projection files that can be input to SMOKE after being reformatted. Control files can be manually created, but for detailed control applications, such as various levels of controls on many different sources applied non-uniformly in a modeling domain, the creation of these files could be unreasonably tedious. We propose the development of a GUI for creating and applying emissions projection and control packets to emissions inventories. Integrating both EGAS control packets and allowing the creation of new projection and control factors, the GUI will display the packets graphically as they apply to different sources and will display inventories before and after the application of projection and controls. This tool would also allow the selection of spatial units (state and/or county) and sources (by SCC) and the level of control to apply to them. The output would be a set of SMOKE-formatted control packets.

Jeff Vukovich, Alison Eyth, Catherine Seppanen, Zac Adelman

Contents