Jim Carton
7-17-2020

I downloaded three monthly discharge datasets: 
1) the latest version of Dai and Trenberth (ucar: s551.0 | DOI: 10.5065/D6V69H1T) 
which includes 925 major rivers through 2018.
2) Seven  major Arctic rivers (greatarcticrivers.org) which goes through early 2020, but I 
truncated 12/2019.
3) Greenland discharge at 267 locations beginning 1/1987 (from Mankoff et al.)

The data was combined using the matlab routine 
/aosc/horse/carton/rivers/data/gridded_discharge_mo_360x720x480.m
and the individual basin discharges were inflated using grads:
/aosc/horse/carton/rivers/data/gridded_discharge_mo_360x720x480.exec
to create the combined 1/2-degx1/2degx480mo gridded nc4 data set: 
/aosc/horse/carton/rivers/data/gridded_discharge_mo_360x720x480_corrected2.nc

In the combination process (see gr...m) the transports of the Arctic rivers was increased by 30% 
as recommended by the greatarcticrivers.org site  to account for ungaged discharge, and any missing values 
were filled in with climatology.  The other rivers were inflated by basin-dependent factors given in 
Dai and Trenberth (2002) Table 3 and then further inflated by 6% to bring the global average discharge up to
1.14Sv.  I note that this is still a bit less than the value of 1.18Sv quoted in Dai and Trenberth. This last step 
was carried out in the grads script (the variable is: discharge2.  Note: longitude is defined 0-360). 

I've included the .m and .exec files in this directory