I noticed the 2020 version of NEI uses surrogates 135 and 136 for the RWC sector instead of 300 (NLCD low intensity). I haven’t been able to find any documentation on the switch, which has large implications. Does anyone know where I can find this information?
The description of the surrogates used in the 2020 platform is in section 3.4 of the 2020 technical support document:
And in this presentation from the 2023 Emissions Inventory Conference:
Hi James, thanks for your response. I’ve looked through both of these files, but I was hoping to find documentation justifying why the change in surrogates was made for RWC from 300 to 135 & 136. From my own research, it seems like this documentation may not exist. Thank you for your help!
Here is some additional information:
When developing modeling platforms, EPA routinely updates surrogates to utilize updated versions of the underlying surrogate databases or to use a different source of data when it is deemed more representative for a particular source category. As EPA was updating from the 2011 National Land Cover Database (NLCD), used for the 2018 and 2019 modeling platform, to the 2019 National Land Cover Database, starting with the 2020 modeling platform, EPA also examined the Residential Wood Combustion (RWC) sector that utilizes the NLCD. This was done to see if there are other sources of spatial data that could improve the geographic representation of the RWC sector when disaggregating the county-level emissions provided by the emissions inventory. For the RWC sector, the EPA was previously using (prior to the 2020 platform) surrogate 300, “NLDC Low Intensity development” which is land cover characterized by 20-49% impervious surface. We initially selected NLCD Low intensity development for RWC to capture geographic areas where there may be houses but generally in less developed spaces. However, this surrogate does not differentiate by development or structure type. The result is that RWC emissions could end up concentrated over roads, commercial, and other low-moderately developed grid cells. EPA thus embarked on changing this surrogate (in the 2020 Platform) to utilize certain housing data provided by the American Community Survey (ACS). The particular attributes are: single family detached, single family attached, dual family and mobile home and combinations of these, depending on the particular RWC specific source category. The use of certain types of housing seemed more reflective of where this type of source would be located. The downside of the ACS housing data is that the census shapes are broad (particularly in rural areas) so the emissions may appear smeared in some plots. When comparing the two approaches (NLCD vs ACS), we found that the ACS approach looked reasonable, and in fact better than the NLCD Low intensity development surrogate. A comparison of the PM2.5 emissions gridded with each of these approaches is shown in the figures below. In the future, we would like to further improve the surrogates due to the resolution of the ACS data (census shapes), and will be examining the use of building structure data weighted by the ACS for future platform updates.
If the pictures don’t come through, you may email me at eyth.alison at epa
2 Likes
Hi Alison,
Thank you for your response and all this information. Were any techniques used to validate the surrogates to any other sources of data (such as source apportionment or local monitoring campaigns)?
Kyan
We are not aware of the use of source apportionment studies or local monitoring campaigns to assess surrogates. While it may be feasible, there are a lot of parameters that could impact the emissions spatially for this sector such as the emission factors for the various appliances (e.g., wood stoves, pellet stoves, hydronic heaters, fireplaces), the amount of wood burned in them, and the appliance distributions. There is also the meteorological data and resolution of the modeling. We are aware that Ben Murphy of ORD is conducting research using measurement campaigns and CMAQ to assess RWC but not with a surrogate focus. With respect to assessing surrogates, we do assess them by creating maps, and sometimes zoom into areas to do visual comparisons. We did this to compare ACS with a data source of building structures in parts of California; this led to our desire to further pursue available building footprint data to develop a building structure-based surrogate