InMap "AllCause" reporting a mortality rate of over 4 trillion

[Natalia Gonzalez]

For all scenarios I ran in InMAP, there is one polygon with an "all cause" of over 4 trillion. This causes the result to show over 27 million deaths in that cell, even though there are only 16k people in the cell. This is clearly an error, but I am not sure where it comes from or how to fix it. The total PM2.5 calculated for that cell is only 0.00501117258136, which is not even the largest concentration in the output shapefile. Has anyone else experienced this and have any advice for me?

Thank you,

Natalia

[Anil Gogebakan]

Sounds like a calculation mistake! Two things come to my mind:

1) Due to misspecification (or misprojection) of the coordinate reference system, calculations are blown up 

2) More likely that you have a simpler calculation mistake around the lines of  output_variables = { 'TotalPM25':'PrimaryPM25 + pNH4 + pSO4 + pNO3 + SOA', 'deathsK':'(exp(log(1.06)/10 * TotalPM25) - 1) * TotalPop * 1.0465819687408728 * MortalityRate / 100000 * 1.025229357798165', 'deathsL':'(exp(log(1.14)/10 * TotalPM25) - 1) * TotalPop * 1.0465819687408728 * MortalityRate / 100000 * 1.025229357798165', }"

As these lines include taking exponential, logarithm, and multiplication and also there is a division in the end, if you messed up in these lines it might cause such a big number. 

Good luck!

Anil

[Natalia Gonzalez]

The strange thing is, all other rows are reasonable. There is only 1 outlier that has deaths > # of people in the cell. So I do not think it is a configuration file issue

[Natalia Gonzalez]

How can I specify the CRS? I used inputs with CRS EPSG:4326. I assumed that this is the default for InMAP since the examples seem to be in this CRS

[Anil Gogebakan]

Then I'd check if it is in the boundary region of your geographic specification as it may be the case that some functions may misbehave at the boundary or on some erroneous projections. That would bring us back to point 1) I would explore coordinate reference system specifications to see if there is something wrong going on.

Best,

Anil

[Natalia Gonzalez]

Sorry, I meant that I used EPSG:4269. Can you please let me know how to specify this in the Config? Thanks for all your help!

[Anil Gogebakan]

That is also the CRS I use often, but after that step, you need to project your map into Lambert Conformal Conic projection so it is a 2-dimensional space that you could work. I would suggest checking the step where you define your projection and make sure this projection also covers the area that you want to cover. Here is an example of my parameters (but this will not be useful for you as you need to find parameters of your own geographic region): 

+proj=lcc +lat_1=40.66666667 +lat_2=41.03333333 +lat_0=41.75 +lon_0=-71.25 +x_0=0 +y_0=0 +a=6370997.000000 +b=6370997.000000 +to_meter=1", inplace=True

Frankly, I don't know why you have this big number problem. But I would start exploring here and then move forward with other potential causes.

Best,

Anil

[Natalia Gonzalez]

Thank you so much!