Basic Holdings
The Global River Discharge Database contains 949 individual listings
that can be found distributed through the six continent-specific
volumes. As was the original intent of the UNESCO series, geographic
coverage is widely distributed at the global as well as continental
scales (Index Maps 1 through 6, below). Separate data volumes are
given for Africa, Asia, Europe, North America, South America, and
Oceania. The holdings are as follows:
Continent Number of Sites
Africa 198
Asia 215
Europe 152
North America 251
South America 104
Oceania 29
Data for Europe considers the area that is generally west of the Ural Mountains and north of the Caucasus. Thus, many sites within Russia and the former Soviet Union are distributed between Europe and Asia. Asia includes the Near and Middle East starting just east of the Red Sea. Its western boundary also includes Turkey. It progresses eastward to include South Asia, Siberia, and the Far East. Mainland Southeast Asia and the Philippines are included. North America includes data within Central America southward to Panama. Oceania we define as Australia , New Zealand, Indonesia, Kalimantan, Sulawesi, New Guinea, and the remaining islands of the maritime continent.
The period of record for the reported discharge values varies widely from station to station. The range is from less than 1 to 178 years, with a mean of 19.3 years. Figure 1 gives a frequency distribution plot of the length of record (in years) for data maintained in RivDIS v1.0. It is clear that the bulk of the data spans but a short period of record. Further, a time series (Figure 2) shows the enormous increase in the number of stations reporting as part of International Hydrological Decade activities, and the sharp decline thereafter. Figure 3 and Figure 4 show time series of drainage area and freshwater discharge of the continental land mass using RivDIS data entries. These plots show the coverage afforded by the most downstream of monitoring stations in all drainage basins represented in the RivDIS data archive. The discharge monitoring stations collectively represent a substantial fraction of continental area and discharge only from the late 1960's and are sensitive to the inclusion or exclusion of individual sites (e.g. inclusion of Óbidos on the Amazon river in the late 1920's folllowed by temporary exclusion just prior to 1950). The best representation was achieved in the mid to late 1970's when the monitoring data represented approximately 40% of global land area (excluding Antarctica) and nearly 50% of global freshwater discharge to oceans, coastal seas and inland receiving waters.
For users interested in determining climatic averages or trends over the last 30 years, we have provided time series plots for each station, as an easy means of visualizing the consequences of including information that may be available for only brief observational periods. Activities associated with programs such as the Flow Regimes from International Experimental and Network Data (FRIEND) and the World Meteorological Programme's Global Runoff Data Center (GRDC) lend hope that a more contiguous global record can be maintained in the future and thereby lessen the impact of historical data losses.
Using the Published Database
Discharge data contained within RivDIS v1.0 are presented for
individual sites. This station data is given in alphabetical order,
by site name, and can be found within each of the continent-specific
volumes. A set of indexes is also presented near the rear of each
volume to assist users in navigating the published data set. These
indexes are sorted by several of the basic data entries within the
archive. Sorts are presented by: country, river name, "discharge
into" (receiving waters specified by UNESCO), major drainage
basin (Figure 5), and RivDIS map code. The map
codes are unique by continent and should facilitate quick location of
individual sites. Maps of the sites for each continent are given in
the index maps in Volume 0 and adjacent to the detailed maps found at
the back of each of the six continent volumes. An abbreviated index
containing some of the information in the individual continent volumes
is given as the "Summary Listing of Sites" in Volume 0.
Index maps in this Volume give a synoptic view of the data holdings
for each geographic region.
Each data entry provides textual information, tabular data, and graphs. Data for a typical site is shown in Figure 6. The top right hand corner shows the continent in which the station resides. On the left, the station name is given followed by data fields for Country, River, "Discharge into", and Major Basin. On the top right is the MapCODE. These data fields are identical to those used in the indexes found near the end of Volumes 1 through 6. Also on the top right are the latitude / longitude coordinates (degrees and minutes; "-" denotes West or South), elevation of the site, and drainage area. Just below are the start / end dates for the record as well as the total number of months contained in the database for this site. Note that there may be discontinuities in the time series. The "Total Months" entry will reflect the true number of individual monthly observations. A small table showing mean annual discharge and runoff computed from the entire observational record is presented. The data are given in units of m3 second-1 and km3 yr-1 and, when area is available from the original UNESCO archives, l sec-1 km-2 and mm yr-1. The monthly analogue of this table is also given. Note that for this particular site, monthly and annual means were computed using data from 1893 to 1986. A plot of mean monthly discharge is also given, again based on the entire period of record. A time series plot of observed discharge is displayed for the period 1960-1990. Superimposed on this plot is a set of lines representing the proportion of observations below the 90th, 50th, and 10th percentile levels. These were calculated from raw data drawn from the entire period of record. The resulting flow duration series is shown in the middle, right-hand portion of the page.
Error Checking
There were two classes of checks instituted in the preparation of
RivDIS v1.0. The first involved the attribute data associated with
each site, while the second involved discharge entries themselves.
The reported site attributes were checked for consistency through
comparison with the UNESCO Discharge of Selected Rivers of the
World published series. Specifically, checks were made on the
accuracy of site names, geographic referencing of reported site
locations, and contributing drainage areas. A GIS-based display of
the site coordinates was superimposed on 1:3 M scale digital line
segments depicting the world's river systems (ArcWorld / ESRI,
Redlands, California) to check for proper positioning. We also
checked positioning using a Simulated Network Topology at 30-minute
resolution (STN-30) that was developed using GHAAS
(Figure 5). Using STN-30, we also compared reported contributing
areas, discharges, and runoff to those simulated by our GIS and
modeling-based system. Those sites yielding inconsistencies were
examined more closely and compared to information provided in
additional independent sources such as standard atlases. Appendix 1
in Volume 0 gives a listing of sites where longitudes and latitudes
were not given, necessitating assignment using UNESCO GeoRep Codes or
Atlas information
The second class of potential correction involved the flow data itself. Several steps were involved: a). an independent comparison of keypunched entries to published values (when available) was first made, b). examination of extreme or unreasonable values for runoff, computed as reported discharge divided by reported area (which itself may have been corrected), and c). a visual examination of the time series plots with a search for outliers. The discovery of any inconsistencies were grounds for a more careful analysis of particularly suspect sites and suitable corrections were then attempted.
If attempts to reconcile the data with independent sources failed, a particular site would be excluded from the database. In addition, there were 77 sites for which the original UNESCO discharge publications gave descriptive information but no reported discharges. A total of 22 stations had discharge data but insufficient information for proper geographic referencing. These were also removed from RivDIS. Those sites which were omitted are given in the Appendix 2 of Volume 0.
