Statistics Service Frequently Asked Questions

Why is only tab-delimited (RDB) format available?

The USGS is starting with a commonly used internal format for serving statistical data for this service. Schemas need to be created to represent statistical water data in XML and JSON. When these are completed they will be integrated into the service.

Why are there no statistics served for discrete sites?

Discrete sites are those where human (manual) measurements are made. However, to generate a statistic it must be based on looking at many values over a time period. Automated time-series sites perform this function, making statistics like minimum, maximum and mean easy to calculate. Months or years may go by before a human visits the site to take a manual measurement. Consequently, a site must serve daily values for statistics to be generated.

Why are provisional data not included in statistics?

Provisional data, as the name implies, is subject to change. USGS hydrologists frequently change provisional data, for example to remove aberrant spikes or to estimate flow when a site is ice affected. After these changes, the data are marked as approved and statistics can be calculated from these values. Since some values will be incorrect or missing if marked provisional, statistics based on provisional data will be misleading, so they are not allowed.

Why are there no statistics for a period because only a few daily values are missing?

To be faithful to the intent of a statistic, it should be calculated based on a complete set of values. For example, a monthly statistics for a given month and year should be based on accurate daily values for all days in the month for the year. Consequently, if there is not a complete record of daily values for that month, statistics for that month are not served. However, the service allows this to be overridden by supplying the missingData argument.

How are statistics calculated? Are they rounded? What rules apply?

See this topic in the USGS Water Data for the Nation help system.