I’m trying to process daily data of Fire emissions in SMOKE4.5 and I obtained this error message in the smkinven.log file
*** ERROR ABORT in subroutine GETPDINFO
No day -specific sources matched the inventory for the time period processed.

The day of analysis is 2017, January 1st (01/01/17)
Attached you will find the FIREEMIS file (Emisiones_incendiosDominio01.csv), the ORL FIRE file (Incendios_Dominio01.txt), the ASSIGN file, the run file (run_point.Incendio_Chile.txt) and the smkinven log file.

I don´t know why smkinven does not process the daily data and the PTDAY file is not generated.
Hope you can help to to solve this issue.
Kind regards

ASSIGNS.Incendios_Chile.txt (23.7 KB)
run_point.Incendio_Chile.csh (9.8 KB)
smkinven.point.incendioD1.txt (20.1 KB)
Emisiones_incendiosDominio01.csv (722 Bytes)
Incendios_Dominio01.csv (380 Bytes)

At first glance, it looks like LOCID needs to be set to 1 in the PTDAY file. Also you have blank line after #DESC line which I don’t think is needed. Typically, we put at least one day of data before the start of simulation and one after so we make sure we have data for desired time zone. So I would put data in for 12/31/2016 and 01/02/2017. I am not sure if time zone is important for your task or not.

1 Like

Hello @jvukovich . Thanks for your reply.

I removed the blank line you mentioned and set to 1 in LOCID column.
Also, I added data to previous and after day and the same error is still appearing.
I don´t know why this warning “WARNING: Period-specific record does not match inventory sources:” is shown.

Attached you fill see the log file
smkinven.point.incendioD1new.txt (20.1 KB)


There’s a warning the log that may be the issue:

—>> WARNING in subroutine STR2REAL
Error reading REAL from " "; IOSTAT= 59

There may be some strange characters in your csv files if you are taking files from Windows apps (e.g. Excel) and trying to use them in a Linux environment. Try using a command like “dos2unix” to get rid of these strange characters in your file that you physically can’t see but are there.
Typically use of dos2unix is like this:


You would use the outfile in the SMOKE SMKINVEN processing.

The problem is the same. I applied dos2unix and the same Warning message is present in the log file. I think that is not the problem.
I think the error is related to YEAR4_BASE and YEAR4_PIVOT variables. I defined in the run file but, there is no information about in the manual of SMOKE.
I need an example of other runs using PTDAY emissions to compare with my data.
Thanks by the way.

We typically use the FF10 format for fires. Some sample files are located here:


Thanks again @jvukovich
I read your files and saw you considered the unit of 1 for these variables

Did you calculate the plume rise in SMOKE using those values?
I am using ORL format according to the manual of SMOKE, because the plume rise calculation in it required that format. However, I would try with FF10_point.

We do use a fire emissions model (Bluesky Pipeline) that provides the heat flux for each fire. We use SMOKE LAYPOINT to calculate the plume rise based on the heat flux. I believe there is some documentation in SMOKE Users Guide about how to roughly estimate heat flux.