Written: August 9, 2018
Revised: Aug 24, 2019 to describe shapefile locations.
Revised: Aug 28, 2018 to describe updates to split AOLGA and AOLGB
Revised: Sep 18, 2018 to descripe updates for the 1KM results
Revised: Nov 06, 2018 to describe updates for the 24KM results
Revised: Nov 12, 2018 to describe updates for the 24Km results
Revised: Dec 1, 2018 to modify the method used to calculate PM0.1 concs

This readme file describes the contents of the updated simulations for 
UFP carried out for 2000-2011.  

Note: The data is considered preliminary until
the publications describing it have been finalized, but an early version
of the information is posted to support collaborating research projects.

Dec 1, 2018 Notes:
===================
The bart5a-4km.zip file contains results that have been updated to 
remove nitrate, sulfate, and ammonium from the first 5 size bins of the
particle distribution and shift the particle radius downward as a result.
The boundary conditions fed into the calculation put too much sulfate in
the PM0.1 which attracted considerable nitrate condensation leading to
particle growth above the 100nm threshhold.  This causes the PM0.1 EC
field to look strange which impacts the epi results.  In reality, the 
background sulfate should be in the accumulation mode which attracts
nitrate into this size range (not into the UFP range).  This correction
was performed as a postprocessing step. 

Nov 12, 2018 Notes:
===================
The bart4a.zip file contains 24KM results that have been updated to
split the ALXY variable into ALXY and ABIO2.  The new ALXY variable
is only from anthropogenic sources.  The ABIO2 variable is dominated
by biogenic contributions but it may have a minor amount from background
sources as well.  

The anthropogenic SOA can be constructed as
 AALK+AXYL+ATOL+ABNZ+AOLGA
The biogenic SOA can be constructed as
 ABIO+ABIO2+AOLGB

Note that the semi-volatile SOA speces were allowed to further react
with OH in this model calculation which increased the predicted SOA
concentrations relative to what was predicted by Hu et al. 2017.

Nov 06, 2018 Notes:
===================
The files under data/soa_v2/monthly_avg3/ describe monthly average
concentrations that have been re-run to fix several errors found in the
previous files loaded to monthly_avg2.  All older files should be deleted
and not used for analysis.  The first set of results posted to monthly_avg3
are for 24KM resolution.  They can be used with the averge-24km.exe program
included inn the zip file bart4.zip.  The basic map projection for these 
files are

CA DOMAIN: XORIG YORIG DX DY: -465.0, -475.0, 24000.0, 24000.0

The standard Lambert Conformal projection is described in the top-level
readme.txt file.  The max number of x,y cells in the domain is

CA: 43,44

Additional updates will be provided as the 4km and 1km results become available.

===============================================================================
*********************************************************************************
OLD NOTES PROVIDED BELOW FOR REFERENCE ONLY.  DO *NOT* USE THE monthly_avg2 FILES.
*********************************************************************************

The files under data/soa_v2/monthly_avg2/ describe monthly average 
concentrations in the different size fractions. The 4km results for the years 
2000-2011 are in the file bart3.zip, in two domains with the same
map projection as described in the original readme.txt file in this 
directory.  The shapefiles grid_scb4km.* and grid_sjv4km_v2.* in the
shapefiles sub-directory align with these data. The 1km results for the
years 2000-2002 are in the file 1KM.zip in three domains.  The map 
projections are the following:

LA DOMAIN : XORIG YORIG DX DY: 167.0,-343.0,1000.0,1000.0
SFBA DOMAIN:XORIG YORIG DX DY: -189.0,25.0,1000.0,1000.0
SAC DOMAIN:XORIG YORIG DX DY: -113.0,145.0,1000.0,1000.0

All domains use the Lambert Conformal projection described for the 
parent 4km results.  The max number of x,y cells in each domain is

SFBA: 100,108 
SAC: 80,64 
LA: 136,88 

A program average.exe for windows 10 is provided to help researchers
use the 4km data files, and a program average1km.exe is provided to use
the 1km data files.  The program should be run in the same directory
where the downloaded data files are extracted.  The input file to the 
program is a comman separated file with a header line #1 containing the 
following field names:

idnumber - integer number assigned to each participant
year.in - the year the participant moved to this location
month.in - the month the participant moved to this location
year.out - the year the participant left this location
month.out - the month the participant left this location
latitude - the latitude of the participant location
longitude - the longitude of the participant location

The program will calcualte the average pollutant concentrations for
the participant and write them to standard output.  A typical debug
sequence to test the program would be

average.exe < participant_info.csv

The results will print to the screen to help understand if any of the 
input data falls outside acceptable ranges.  Read the error message if
it crashes.  Contact Mike Kleeman (mjkleeman@ucdavis.edu) if all else
fails (you should really contact him anyway to get some advice on how
to best use the strong aspects of the data and avoid the weak aspects).

A typical production sequence to use the program would be

average.exe < participant_info.csv > results.csv

so that the output goes into a file that can then be used for further
calculations.

Note: an earlier version of the data bart2.zip and average.exe combined
the oligomerized SOA products AOLGA and AOLGB into a single variable.
The current version of the data bart3.zip and updated average.exe
split those SOA products into AOLGA (anthropogenic oligomerized SOA) and
AOLGB (biogenic oligomerized SOA).  Update both the data and the program
to use the new results.

Weekly average results or possibly even daily average results are also
available as needed.  Contact Mike Kleeman (mjkleeman@ucdavis.edu) if
data with higher time resolution is needed.

Michael J. Kleeman Research Group
University of California Davis
mjkleeman@ucdavis.edu