HYSPLIT @ IF-USP
Version of this readme file: 24/July/2018

A few matlab files were corrected. They had julian-date variables with
the wrong dimension. They were written as a matrix (N x N) instead of
(N x 1), hence occupying more space than necessary. There was no
dataloss or problems with the previous version, as the correct jd
could be found at jd(:,1) or jd(1,:).

--------------------------
1/March/2017

We did a special run with 12min output, which matches the time-step of
the SIPAM radar and hence makes it easier to interpolate. Only backward
trajectories and only 48h. However, we did from 2014 to 2017, and the
output tgz are separated by year.

--------------------------
24/August/2015

Forward and Backward trajectories were calculated using the Hysplit
model for all GoAmazon sites (T0a, T0z, T0e, T1, T2 and
T3). Trajectories were started 100m above ground, every 30min between
Jan/1st and Dec/31st 2014. The structure of this repository is:

.
|-- gdas0p5
    |-- backward
    |   |-- ascii_files
    |   `-- matlab_files
    `-- forward
        |-- ascii_files
        `-- matlab_files
`-- gdas1
    |-- backward
    |   |-- ascii_files
    |   `-- matlab_files
    `-- forward
        |-- ascii_files
        `-- matlab_files

Data in folder gdas1/ has the output of the simulations forced by 1 x
1 deg GDAS1/NOAA boundary conditions, while under the gdas0p5/ folder
you will find those using the 0.5 deg files.

Inside the boundary condition folder you will find a forward/ and
backward/ folders... and inside each, a matlab/ and ascii/ folders. 

In the ascii folder you will find the standard Hysplit output
files. See their description at the Hysplit website. For both the
backward and forward files we produced indepent files for each site on
the output.

In the matlab folder, you will find one file for each site. When you
load such a file in matlab, you will find two matrices, for instance:

gdas1_backward_T3[17520, 673, 20]
jd[17520]

Organized as follows:

17520 = number of 30-min time steps in a year
673 = number of 30-min time steps along a 14 day trajectory
20 = output columns from hysplit:

%1-station       %11-lon (o) 
%2-tracer        %12-alt (m) 
%3-YY            %13-PRESSURE
%4-MM            %14-THETA   
%5-DD            %15-AIR_TEMP
%6-hh            %16-RAINFALL
%7-mm            %17-MIXDEPTH
%8-????????      %18-RELHUMID
%9-time(h)       %19-TERR_MSL
%10-lat (o)      %20-SUN_FLUX

Henrique Barbosa
hbarbosa@if.usp.br

Glauber Cirino
glauber.cirino@gmail.com