This forum is shutting down! Please post new discussions at

Interpreting Recreation Photo-user-day results


I’m using the Recreation model for the first time to
estimate public recreational use of a mostly conserved Valley near Seattle, Washington. I’ve run the model
successfully with a mile and half-mile gridded cells (without regression) and am trying to interpret
the results.

I know for a fact that there are thousands of users annually
on the Mount Si trail, but the model is only showing the average annual
photo-user-days as closer to 15. Also, many other locations in the Middle Fork
Snoqualmie watershed that are popular destinations are showing 0 PUDs.

Wood et al 2013 suggests incorporating additional
socioeconomic factors to explain local variability. Is that what I need to pursue to improve the results? Or are there other ways I can derive actual user
days for this location from the model?

Any suggestions very welcome!



  • DaveDave Member, Administrator, NatCap Staff
    Hi @BenHughey,

    The relationship between actual number of annual visitors and annual photo-user-days is different in different places. So it's hard to know how many actual visitors 15 photo-user-days represents without having some baseline visitor data. If you have some visitor counts for certain areas, you could use the ratio of PUD to actual visitor counts in those areas, and apply that ratio to other similar places where you don't have counts. We typically use a linear regression for that.You can examine some of the scatterplots in Wood 2013 to see how PUDs
    relate to actual visitor counts at the sites in that study.

    Another way to improve accuracy of counts is to experiment with drawing polygons that represent the specific areas/trails you are interested in, and then running the model without grid cells. That way you get to directly define the geographic area at which photo-user-days get summarized. Your AOI shapefile may contain multiple polygons.

    That excerpt from Wood 2013 suggests that a variety of factors about the natural and built environment, population demographics, etc are all at play in explaining the variability of visitation across different sites. So that's more about computing a regression to understand which attributes of the study area are driving visitation patterns.

  • Thanks for the thorough and speedy response Dave!
    Sounds like I need to read up more on using the linear regression and find a source for visitor information. 
    I appreciate the help!

Sign In or Register to comment.