Why are Eco & Streets estimates different for the same trees

Section for FAQs regarding i-Tree Eco (UFORE)

Moderators: i-Tree Support, i-Tree Team

Post Reply
i-Tree Team
Posts: 365
Joined: Fri Jun 27, 2008 10:50 am

Why are Eco & Streets estimates different for the same trees

Post by azelaya » Thu Feb 09, 2017 6:33 pm

Eco and Streets have completely different underlying methodologies which accounts for much of the differences in the model estimates. This is explained in the documentation that describes the underlying functions, assumptions and limitations of these models in the archives section of the website. https://www.itreetools.org/resources/archives.php .

There is a lot of information available but for a snapshot of how Streets works see the following 1-pager:

Reference Cities - The Science Behind i-Tree Streets (STRATUM) - A summary document explaining the use of regional reference city data as the basis for i-Tree Streets cost and benefit modeling capabilities.

For further details on Streets methods, you have to refer to the USFS Reference City Research guides. (Accessible from the Archives page also) Below is the Northeast Guide link but others are available for each climate region.

Northeast community tree guide: benefits, costs, and strategic planting. Gen. Tech. Rep. PSW-GTR-202 Albany, CA: U.S. Department of Agriculture, Forest Service, Pacific Southwest Research Station. - [4.24 MB]

Each guide has region specific information and Appendix 3 explains the underlying methods for the region. The Indianapolis guide is one of most recent guides (2009) and mentions model limitations at the end of Appendix3.

Which model is more accurate?
If accuracy of estimates is your primary concern, we recommend using i-Tree Eco and providing all the recommended tree measurements (tree height, canopy width, crown light exposure, % missing, etc.) in the Eco interface. The chart below from the Eco overview page shows how these measurements are used to model a tree’s functional leaf surface area and leaf biomass and how these variables affect various model ecosystem service estimates. Basically, Eco is modeling how a tree’s canopy – based on actual measurements – interacts in local conditions based on hourly weather and hourly pollution interactions, and other environmental factors. The links to the following (2) documents (pdf) describe this more detail.

Data Variables - Ecosystem Services Relationships descriptions - https://www.itreetools.org/eco/resource ... nships.pdf

Eco Data Limitations guide - https://www.itreetools.org/resources/ma ... ations.pdf


Considerations when using Eco with less tree variables
If some recommended tree measurements are not entered in Eco, the model is less accurate because it is then using regression equations to estimate missing data variables such a tree height, canopy width, and using defaults for other variables like crown light exposure. We don’t have information regarding the degree of accuracy lost or a difference in estimates when using Eco with limited data vs all recommended tree variables. The accuracy of estimates will vary by situation and over the size of the tree population being estimated. Estimates and values for a single tree may be very different if it is not typical of tree size characteristics Eco generates based on limited data. However, estimates may be normalized over a population of trees. This is discussed in more detail in the Data Limitations guide link above. Nonetheless, we still recommend using Eco over Streets because the Eco model is updated more frequently and is using local pollution and weather data compared to Streets. Also, you may be able use provide some additional variables like tree height and crown light exposure which will help improve estimates if all are recommended variables are not available.

i-Tree Streets Considerations
The Streets model has not been updated significantly since the original reference city research was completed. Streets will not be updated in the future as it is too costly to update and maintain multiple models and Streets is less applicable for broader use. Therefore, Eco will be the primary model used for ground assessments going forward as the USFS Northern Research is updating and integrating new science in the Eco model continuously. We still support i-Tree Streets users as it is still a very popular tool especially with its primary users group – U.S. Street and Municipal Tree Managers.

We don’t recommend using Streets outside the US because of limitations in applying the reference city research abroad. Yes, people still do use Streets in other countries but they have to recognize the limitations. International i-Tree Streets users accept that their street trees & city, are being modeled as a city in the US and their city has similar climate, street tree growth characteristics, management & pruning practices, power (electric, natural gas, hydro -electric) generation characteristics and usage, and local costs. This regional model approach offers the user easy use and access but has limitations in the US also as cities vary throughout any of the climate regions. Street trees are pruned and cared for differently than yard, park, natural area trees which also limits applying Streets estimates more broadly beyond right of way trees.

Another reason you will see different estimates between Eco and i-Tree Streets is because some Streets estimates are based largely on the averages from the reference city research. For example, important variables like tree condition do not affect ecosystem services estimates in Streets like rainfall interception, carbon sequestration, etc. Also, energy effects are applied as an average for all trees in i-Tree Streets regardless of their proximity to a structure. Conversely, Eco requires multiple variables to model tree/building energy effects more accurately, which may result in negative values.

Here are some quick thoughts and references on other questions related to Eco vs Streets estimate differences based on a user's project comparisons.

1) Annual air quality benefit is much higher for i-Tree streets compared to Eco. Opposite to this, the average benefit pricing is lower for streets comparded to eco: $12.06 vs $17.75/kg. Benefit pricing cannot be adjusted in Eco. Which is best to use?

First, there is a bug in Eco v6.0.3 causing problems with updating benefit prices in Eco which we expect a new release to resolve very soon. Eco is best to use for reasons stated above with all recommended tree measurements preferably. Eco uses more updated air pollution modeling and valuation based on the US EPA BenMAP health impacts model which values air pollutant concentration reduction based on affected human population characteristics. This is adapted differently for countries outside the US. The following documents describes how Eco pollution modeling is adapted.

i-Tree v6 Pollutant Removal, Biogenic Emissions and Hydrologic Processes Descriptions v1.1 - A document (pdf) describing i-Tree Eco v6 ecosystem services methods and processes integrated in i-Tree Landscape, Canopy, Design and Forecast including international adaptations for Australia, Canada and the United Kingdom. (Updated April 14, 2016) https://www.itreetools.org/landscape/re ... ations.pdf

2) Annual stormwater capture is much higher in Streets than Eco. What would cause this?

Different modeling methods and different valuation of avoided run off in Eco vs. rainfall interception in Streets. Eco is using a with-and-without tree simulation based on the i-Tree Hydro model to estimate tree hydrologic effects, which consider above and below ground processes to greater extent to value avoided runoff. The Hydro based model tends to be more conservative than Streets estimates.

See the following (2) docs for more details.
i-Tree Eco Precipitation Interception Model Descriptions A description of the precipitation interception model integrated in the i-Tree Eco version 5 application. (Updated 01/30/13) https://www.itreetools.org/eco/resource ... ptions.pdf

i-Tree Streets/Design/Eco Rainfall Interception Model Comparisons A document discussing the differences in methodologies between the rainfall interception model, used by i-Tree Streets and Design, and the avoided runoff model methodology used in the i-Tree Eco version 5 application. https://www.itreetools.org/eco/resource ... risons.pdf

3) Structural values (replacement cost) are quite different. Do Streets & Eco use different assumptions?

Yes, these are somewhat different methods. The structural estimate is not meant to be exact accounting in Eco in the same way that an arborist would use the CTLA formula to assess onsite costs for replacing an individual tree. The CTLA methodology is adapted in Eco to capture a structural or compensatory resource cost that is not accounted for in the ecosystem services estimates. Some Eco variables are adapted to represent what the CTLA appraisal method uses for estimating replacement cost: tree size, location, condition & species factors. Since Eco has flexible data options, all variables may not be used and defaults may be applied.

Although the following (2) documents are somewhat old, they will give you an idea of the adapted version of CTLA that Eco uses.

A Ground-Based Method of Assessing Urban Forest Structure and Ecosystem Services - A 2008 journal article describing the Urban Forest Effects Model (UFORE) data collection variables, model methods, advantages and limitations. (page 350) https://www.itreetools.org/eco/resources/08 UFORE.pdf

This original model structural value methodology is described in more detail in an earlier document http://www.nrs.fs.fed.us/pubs/jrnl/2002 ... ak_003.pdf

the CTLA adaptation information for Canada and other countries is based on information provided from partnering collaborators in those countries.

Streets relacement value methodolgy can be found at this forum post https://forums.itreetools.org/viewtopic.php?f=20&t=804

4) i-Tree streets reports total carbon dioxide captured, whereas Eco reports total carbon captured. I converted the amount of CO2 to carbon to compare the datasets and the results are quite similar, however the pricing is different. Do the systems use different pricing?

Yes, Eco using more recent pricing based on the Social Cost of Carbon in the US. Streets value will be older based on reference city research.

Social Cost of Carbon for Regulatory Impact Analysis - Under Executive Order 12866 - A technical support document from the Interagency Working Group on Social Cost of Carbon. United States Government. May 2013. A updated 2013 reference document used for i-Tree version 5 & 6 carbon valuation updates.https://www.itreetools.org/eco/resource ... update.pdf

Hope this helps explain some of the model differences and limitations. All i-Tree models have advantages and limitations to consider and we do encourage users to understand these when considering if a model can best meet their objectives.
A member of the i-Tree Team
Post Reply