search

PardeeCenterDU / IFs-Issues-Tracking

This repository only holds the list of bugs that have been reported for IFs. Anyone may add a bug report, but please look to see if your issue has already been added!
2 stars 0 forks source link

ENVPM25 #323

Open jonathandmoyer opened 4 months ago

jonathandmoyer commented 4 months ago

It's an odd variable to change using a multiplier. I'm not exactly sure why, but the multiplier is kind of fickle. I was trying to get our base to match the following graph (the blue line). The base declines pretty significantly and I think that's because all the drivers are positive. I'm not sure why that should be the case based on the relationship between pollution and development, so I'm skeptical about the drivers. We should revisit them at some point. But trying to get a multiplier to flatten our forecast required me to interpolate a value over 15 years to 1.7 or so, then interpolate another value to 5 (the max) over 60 years or so. The downward pressure of the model pushes so hard against it something must be incorrectly specified or overwrought in its implementation.

image

PardeeCenterIFs commented 2 months ago

This is an exponential function, which explains why is so strong. It's also hard coded.

                ForVal = CSng(Math.Exp(5.1238 + 0.2533 * Math.Log(GDPPCP(R%)) - 0.056957 * (Math.Log(GDPPCP(R%))) ^ 2 -
                0.4758 * Math.Log(EDYRSAG25(R%, 3)) - 0.0137 * (BaseYear - 1989 + IY - 1) - 0.14 * GDS(R, Health) / GDP(R) * 100))
jonathandmoyer commented 2 months ago

Wait, the model projection pm2.5 is exponential? Why? That makes absolutely no sense... Let's add this to the modeling pod agenda for when I'm back

PardeeCenterIFs commented 2 months ago

Today during the Modeling Pod we agreed this Exponential function should be changed, and maybe ask the fellows team to look for an alternative.

Changes using the parameter are too sensitive because the function is Exponential.