Revision for Water Data

Surface Water Technical Memo 2017.06, Procedures for Identifying and Documenting Revisions to USGS Water Data

• Groundwater Revisions Guidance
• Surface Water Revisions Guidance
• Water Quality Revisions Guidance

Revisions for Approved Time-series and Discrete Water Data for Groundwater, Surface Water, and Water Quality

Surface Water Technical Memo 2017.07, Procedures for Identifying and Documenting Revisions to USGS Water Data

• Groundwater Revisions Guidance
• Surface Water Revisions Guidance
• Water Quality Revisions Guidance

The guidance below refers to developments in SIMS and NWISWeb that are not yet complete. The SIMS "Revision History" element and NWISWeb revision statement display functions will not be available until the end of the fiscal year 2017. In the interim, revision statements should be added to AQ "Notes" for time-series data, as instructed, and the SIMS Manuscript "Revised Records" element instead of "Revision History". The information will be compiled and displayed on NWISWeb when programming is complete.

Groundwater Revisions Guidance

Contents:

Groundwater: Revision of records due to calibration checks of equipment or instrumentation
Groundwater: Revision of land surface datum (LSD)
Groundwater: Revision of Water Level Records
Groundwater: Example Revision Statements

Groundwater: Revision of records due to calibration checks of equipment or instrumentation

Equipment and instrumentation (water-level tape, pressure gauge, transducer, etc.) calibration checks should be done before the device is used to collect water-level measurements. Prior to calibration, it will be difficult to determine when the device became affected; therefore, it is recommended that the user not adjust any water-level values collected prior to the HIF or WSC calibration check. Known calibration corrections should be applied to the data during collection of the measurement. Users may apply calibration corrections to previously collected data if the calibration correction was due to a factory default error and the device has not been affected by other causes. When a device is re-calibrated then the new calibration log is to be used for all future measurements. Do not prorate across different calibration check logs.

Discrete water-level records must clearly state which device, including equipment identification number, was used at the time of measurement. Revise approved records only when the corrected value differs by more than twice the accuracy of the water level and the above conditions have been met. Continuous records may also need to be revised if discrete water-level values used to apply data corrections are revised.

Scenario 1: A calibrated electric water-level tape was used to measure water-levels in a well but the known tape correction at the time of the measurement was not applied to the data before the water-level record was approved. The accuracy associated with the measurement was 0.01 ft. The tape calibration correction for the measurement was equal to or greater than 0.025 ft.

Steps to take:

1. For approved discrete water-level measurements with a corrected value that exceeds twice the accuracy value: un-approve the measurement, correct the discrete water level, document revisions in SIMS "Revision History" element by describing the change that has occurred and why, and re-approve the measurement.
2. If discrete water-level records at a site are revised, and the continuous water-level records need to be recomputed: Un-approve the affected period of continuous record, reanalyze all aspects of the affected period, mark affected computed water levels with "Revised" qualifier [AQUARIUS], document the revised period with a "Note" [AQUARIUS] and re-approve period.

Scenario 2: A 15 psi transducer was used to measure continuous water-levels in a well. The transducer had a known post factory calibration correction that needed to be applied to the data but the transducer correction was not applied to the data before it was approved. The accuracy associated with the recorded measurements was 0.01 ft. The transducer calibration correction for the measurements was equal to or greater than 0.025 ft.

Steps to take:

1. For approved continuous water-level record in which the revised change exceeds twice the accuracy value: un-approve the affected period, reanalyze all aspects of the affected period, mark affected computed water levels with "Revised" qualifier [AQUARIUS], document the revised period with a "Note" [AQUARIUS] and re-approve period.

Groundwater: Revision of land surface datum (LSD)

Scenario 1: Station levels have been run and a difference between the known benchmark and the land surface datum (LSD) reference mark at the well has changed due to errors in a previous measurement of these reference marks. The difference was equal to or greater than 0.015 ft. According to Techniques and Methods Report 3-A19 (Kenney, 2010), a datum correction is needed. The time period affected by the application of the datum correction contains approved water-level data. This scenario assumes the benchmark and the LSD are stable and have no effect on the change.

Steps to take:

1. Enter revised LSD information into GWSI Sitefile. Create a new record or modify previous record in the GWSI Altitude Datum History table for the active LSD documenting the previous value in the remarks field. Note: multiple LSD records cannot have the same active periods.
2. Create a statement in SIMS "Reference Mark" and/or "Datum" elements that a change has occurred with the LSD and the date of the change. Document changes to LSD in SIMS "Revision History" element.
3. For approved water-level records recorded as water-level altitude, if the change exceeds twice the combined accuracies (altitude and water-level): un-approve the affected period of continuous and discrete water-level measurements, correct the water-level measurements, reanalyze all aspects of the affected period, mark affected computed water levels with "Revised" qualifier [AQUARIUS], document the revised period with a "Note" [AQUARIUS] and re-approve period.
4. Document revisions to discrete water-level measurements in SIMS "Revision History" element by describing the previous value and the reason for the change.

Scenario 2: Station levels have been run between the known benchmark and the land surface reference mark at the well (LSD) resulting in a more accurate measurement of the LSD than the previous value. Difference was equal to or greater than 0.015 ft. According to Techniques and Methods Report 3-A19 (Kenney, 2010), a datum correction is needed. The time period affected by the application of the datum correction contains approved water-level data. This scenario assumes the benchmark and the LSD are stable and have no effect on the change.

Steps to take:

1. Enter new LSD information into GWSI Sitefile. Create record in the GWSI Altitude Datum History table for the previous and new LSD values, dates, method, datum, and reason codes. Note: multiple LSD records cannot have the same active periods.
2. Create a statement in SIMS "Reference Mark" and/or "Datum" elements that a change has occurred with the LSD and the date of the change. Document changes to LSD in SIMS "Revision History" element.
3. For approved water-level records recorded as water-level altitude, if the change exceeds twice the combined accuracies (altitude and water-level): un-approve the affected period of continuous and discrete water-level measurements, correct the water-level measurements, reanalyze all aspects of the affected period, mark affected computed water levels with "Revised" qualifier [AQUARIUS], document the revised period with a "Note" [AQUARIUS] and re-approve period.
4. Document revisions to discrete water-level measurements in SIMS "Revision History" element by describing the previous value and the reason for the change.

Scenario 3: Station levels have been run and a difference between the known benchmark and the altitude of the land surface reference mark at the well (LSD) has changed due to physical changes brought on by subsidence or upheaval of the LSD that was equal to or greater than 0.015 ft. According to Techniques and Methods Report 3-A19 (Kenney, 2010), a datum correction is needed. The time period affected by the application of the datum correction contains approved water-level data. This scenario assumes the benchmark is stable.

Steps to take:

1. Enter new LSD information into GWSI Sitefile. Create record in the GWSI Altitude Datum History table for the previous and new LSD values, dates, method, datum, and reason codes. Note: multiple LSD records cannot have the same active periods.
2. Create a statement in SIMS "Reference Mark" and/or "Datum" elements that a change has occurred with the LSD and the date of the discovery.
3. Future water-level records must use the new LSD altitude.
4. For approved water-level records recorded as water-level altitude, no change is required to the historic water-level record because the actual date the LSD changed is unknown; only the date that the new altitude of the LSD was measured is known.

Scenario 4: The land surface reference mark at the well (LSD) has not changed but a change in the geodetic vertical datum used has occurred resulting in a new value for the LSD.

Steps to take:

1. Enter new LSD information into GWSI Sitefile. Create record in the GWSI Altitude Datum History table for the previous and new LSD values, dates, method, datum, and reason codes. Note: multiple LSD records cannot have the same active periods.
2. Create a statement in SIMS "Reference Mark" and/or "Datum" elements that a change has occurred with the LSD and its geodetic datum and the date of the change.
3. Future water-level records must use the new LSD altitude.
4. For approved water-level records recorded as water-level altitude, no change is required to the historic water-level record because the geodetic datum referenced is part of the water-level record. Station analysis should indicate which geodetic datums were active and when.

Groundwater: Revision of Water Level Records

Scenario 1: Approved water-level record has been found to be associated with the wrong site.

Steps to take:

1. Create new water-level record for the correct well. Review new record and update approval status.
2. Flag original discrete water-level record as "rejected" in GWSI at the erroneous site. Describe in the record level remark field why record is being rejected.
3. Grade [AQUARIUS] original continuous water-level records as "unusable" at the erroneous site. Describe in "Notes" [AQUARIUS] why the record is not usable.
4. For the site in which the water-level record was rejected, create a statement in the SIMS "Revision History" element describing changes in discrete water-level records.

Scenario 2: Approved discrete water-level value has been calculated incorrectly (incorrect hold, cut, equipment, deviation corrections, etc.).

Steps to take:

1. For approved discrete water-level record with a corrected value that exceeds twice the accuracy value: un-approve discrete water level measurements, correct the water-level measurements, document revisions in SIMS "Revision History" element by describing previous value and the reason for the change, and re-approve measurements.
2. If site has continuous water-level records and the discrete water-level records are revised then the continuous water-level records will need to be recomputed with the revised discrete water-level values used to apply data corrections. Un-approve period, reanalyze all aspects of the affected period, mark affected computed water levels with "Revised" qualifier [AQUARIUS], document the revised period with a "Note" [AQUARIUS] and re-approve period.

Scenario 3: Approved water-level record has an incorrectly documented Measuring Point (MP) correction length.

Steps to take:

1. Modify the affected MP record in GWSI with the proper correction length.
2. Create a statement in SIMS "Reference Mark" and/or "Datum" elements that a change has occurred with the MP and the date of the change. Document changes to MP in SIMS "Revision History" element.
3. For approved water-level record recorded as depth below LSD in which the revised change exceeds twice the combined accuracies (MP and water-level): correct the water-level measurements, document revisions in SIMS "Revision History" element by describing previous value and the reason for the change, and re-approve measurements.
4. For approved water-level record recorded as water-level altitude that used the MP correction length as part of its final computation in which the revised change exceeds twice the combined accuracies (altitude, MP, and water-level) correct the water-level measurements, document revisions in SIMS "Revision History" element by describing the previous value and the reason for the change, and re-approve measurements.
5. If site has continuous water-level records and the discrete water-level records are revised then the continuous water-level records will need to be recomputed with the revised discrete water-level values used to apply data corrections. Un-approve period, reanalyze all aspects of the affected period, mark affected computed water levels with "Revised" qualifier [AQUARIUS], document the revised period with a "Note" [AQUARIUS] and re-approve period.

Scenario 4: Approved water-level record has been found to be questionable. Records do not make hydrologic sense. Reason may or may not be determined and the record cannot be corrected.

Steps to take:

1. Flag discrete water-level record as "rejected" in GWSI. Describe in the record level remark field why record is being rejected.
2. Grade [AQUARIUS] original continuous water-level records as "unusable". Describe in "Notes" [AQUARIUS] why the record is not usable.
3. Document rejection in SIMS "Revision History" element describing rejected water-levels.

Scenario 5: Approved water-level record has been found to have erroneous or missing meta-data values such as source, source agency, status, method, etc.

Steps to take:

1. No changes to water-level record metadata fields are required for values set to "Unknown/Transferred" by the database version scripts. These "transferred" values are considered acceptable. For those same records that are later updated from "Unknown/Transferred" with true values, no revision statements are needed in the record level remark field or in the SIMS "Revision History" element.
2. Correct only the approved water-level record meta-data values as needed that have an impact on how the data are used or stored. Describe in the record-level remark field the change that has occurred and why.
3. No revision statement needed in SIMS "Revision History" for these changes.

Groundwater: Example Revision Statements

As outlined in the guidance presented above, revisions to time-series data are to be documented in AQUARIUS using a "Note" in the Data Corrections Toolbox, and revisions to other data types are to be documented in the "Revisions History" element in SIMS. Below are some example statements to be used in either the "Note" or "Revision History" element. These example statements contain four fundamental pieces of information that all revision statements must have:

• What parameter was revised?
• What period was revised?
• When was the revision made?
• Why was the revision made?

The examples below are meant to be used to guide the development of any groundwater data revision statements and all statements must begin with "Revisions:"

Examples:

Revisions: Groundwater levels above NGVD 1929 for the period Nov. 3, 2014, at 08:00 to Jul. 12, 2015, at 14:00 were revised on Dec. 22, 2015, based on changes to the land surface datum.

Revisions: Groundwater levels for the period Aug. 1, 2013, to Dec. 5, 2015, were revised on Jan. 10, 2016, because Measuring Point (MP) 1 correction length was incorrect. Previous MP correction length was 1.15 ft above LSD. New corrected MP length is 1.51 ft above LSD.

Surface Water Revisions Guidance

Contents:

Surface Water: Drainage Area Revisions
Surface Water: Historic Daily and Monthly Discharge Statistic Revisions [with no historical Instantaneous Values in NWIS]
Surface Water: Revision of records due to change found in reference gage from levels
Surface Water: Revisions to Approved Stage
Surface Water: Discharge measurement revisions [gage height or discharge]
Surface Water: Revision due to change in stage-discharge rating or shift
Surface Water: Revision due to change in velocity-velocity and/or stage-area rating for an Index-Velocity Site
Surface Water: Peak Extremes [gage height and discharge] Revisions
Surface Water: Revisions in precipitation data
Surface Water: Reanalysis
Example Revision Statements

Surface Water: Drainage Area Revisions

Scenario 1: Drainage Area (DA) is found to be in error by more than 10 percent.

Steps to take:

1. Correction to the SiteFile in Ground-Water Site Inventory (GWSI) is required if the error in drainage area is greater than 10 percent.
2. Document revisions in SIMS "Revision History" element.

Scenario 2: Drainage Area is found to be in error by more than 2 percent but less than 10 percent.

Steps to take:

1. Correction to the SiteFile in GWSI should be considered if the error in drainage area is between 2 and 10 percent.
2. Document revisions in SIMS "Revision History" element.

Scenario 3: Drainage Area is found to be in error by less than 2 percent.

Steps to take:

1. Correction to the SiteFile in GWSI is allowed if the error in drainage area is less than 2 percent. Several reasons exist to justify such a change:
   
   • Change is justified by data and need
   • Change in DA of upstream or downstream site creates a conflict.
   • Change required to match the DA of a cooperating agency.
2. Document revisions in SIMS "Revision History" element.

Note: If a revision to DA is performed, the significant digits of the DA should still conform to previous reporting rule:

"Drainage areas should be published to the nearest square mile for areas greater than or equal to 100 square miles (mi²); to the nearest tenth of a square mile from 10 to 100 mi², and to the nearest hundredth of a square mile for areas less than 10 mi²."

Surface Water: Historic Daily and Monthly Discharge Statistic Revisions [with no historical Instantaneous Values in NWIS]

Scenario 1: Historical daily and/or monthly discharge statistics are found to be in error and no instantaneous values (IVs) exist to correct.

Steps to take:

1. In the absence of IV data, daily and monthly discharge statistics must be revised only if a monthly mean discharge is in error by 10 percent or more, or if a daily discharge is in error by 50 percent or more.
2. If the selected daily discharge statistics in a given water year (WY) need to be revised, revise as many days necessary to avoid sharp breaks in the hydrograph at the beginning and end of the revised period.
3. If any daily discharge statistics are revised, the corresponding monthly and yearly discharge statistics based on the new daily figures also must be revised.
4. Mark all revised statistics with a "Revised" qualifier [AQUARIUS], document the revised period with a "Note" [AQUARIUS] and re-approve period.

Surface Water: Revision of records due to change found in reference gage from levels

Scenario 1: Station levels have been run and a difference between the reference gage and gage datum that was equal to or greater than 0.015 ft was found. According to Techniques and Methods Report 3-A19 (Kenney, 2010) a datum correction is needed. The time period affected by the application of the datum correction contains approved data.

Steps to take:

1. If the determined datum correction is to be prorated because the cause for the difference occurred over a period of time and not instantaneously, compute prorated correction for each site visit to determine if reference gage readings associated with site visits in approved period need to be corrected (i.e. equal to or greater than 0.015 ft) using Proration_Comp_Tool_V1_1C.xlsx. For any approved period where prorated corrections to site visit reference gage readings are less than 0.015 ft, no revisions are needed and a prorated datum correction, if applied in the unapproved period, should start at the beginning of unapproved period with a 0 correction value. (do not go on to step 2)
2. Revise all site visit reference gage readings during the affected approved time period by applying the determined correction values that are equal to or greater than 0.015 ft to all reference gage readings. If proration is being used (computed in step one), use prorated correction values obtained with the Proration_Comp_Tool_V1_1C.xlsx tool, to revise reference gage readings.
3. For each site visit affected by the datum correction, use the revised reference gage readings to recompute and revise the gage height assigned to the measured discharge. (see "Surface Water: Discharge measurement revisions" section)
4. Document revisions to gage heights assigned to measured discharges in SIMS "Revision History" element.
5. For each site visit affected by the datum correction, compare new reference gage readings to recorder readings.
6. If recorder readings, which have an accepted uncertainty of 0.02 ft, differ from the new (corrected) reference gage readings by 0.04 ft or less, no revisions of recorded gage heights are needed. (do not go on to step 7)
7. If recorder readings differ from the new (corrected) reference gage readings by more than 0.04 ft during approved period, un-approve period, reanalyze (see "Surface Water: Reanalysis" section) all aspects of the affected period, mark affected gage heights with "Revised" qualifier [AQUARIUS], document the revised period with a "Note" [AQUARIUS] and re-approve period. (do not go on to step 8).
8. For each site visit affected by the datum correction, compute the percent difference between measured discharge and computed discharge using the revised measurement gage height and the approved shifted rating at the time of the measurement.
9. If the computed percent difference is less than two times the assigned uncertainty of the discharge measurements no revisions of computed discharge are needed. (do not go on to step 10)
10. If any of the computed percent differences are greater than two times the assigned uncertainty of the discharge measurement, un-approve period, reanalyze (see "Surface Water: Reanalysis" section) all aspects of the affected period, mark affected computed discharges with "Revised" qualifier [AQUARIUS, document the revised period with a "Note" [AQUARIUS] and re-approve period.

Proration Worksheet

Surface Water: Revisions to Approved Stage

Scenario 1: An error has been found in an approved stage value(s).

Steps to take:

1. If stage values, which have an accepted uncertainty of 0.02 ft, has been found to be in error by 0.04 ft or less, no revisions of gage heights are required.
2. If published stage values have been found to be in error by more than 0.04 ft during approved period, un-approve period, reanalyze (see "Surface Water: Reanalysis" section) all aspects of the affected period, mark affected gage heights with "Revised" qualifier [AQUARIUS], document the revised period with a "Note" [AQUARIUS], and re-approve period.

Surface Water: Discharge measurement revisions [gage height or discharge]

For site visit based revisions, the gage height associated with the measured discharge and the measured discharge are considered to be the only primary data; other information stored in the database associated with site visits is considered to be ancillary. No specific revision criteria or revision documentation requirements are provided for ancillary data. WSCs are encouraged to make changes to ancillary data when an error is found and considered to be significant.

Scenario 1: The gage height associated with an approved discharge measurement is found to be in error.

Steps to take:

1. If the gage height associated with an approved discharge measurement is found to be in error by more than the assigned uncertainty (as determined directly by the hydrographer) of the reference gage readings that were used to determine the mean gage height, un-approve the measurement, revise the gage height value, and re-approve the measurement. Note: The criterion for a revision of the gage height is associated directly with uncertainty observations made by the hydrographer.
2. Document revisions in SIMS "Revision History" element.
3. Evaluate need for revision of rating or shift (See "Surface Water: Revision due to change in stage-discharge rating or shift" section).
4. Evaluate need for revision of peak-flow file (See "Surface Water: Peak Extremes Revision" section).

Scenario 2: The discharge associated with a discharge measurement is found to be in error. For example, an error in processing of a discharge measurement results in a change in the final total discharge for the measurement.

Steps to take:

1. If the discharge associated with an approved discharge measurement is found to be in error by more than the assigned uncertainty (as assigned directly by the hydrographer based upon available information, including, but not limited to, a qualitative assessment of measurement conditions, uncertainty estimates provided by QRev for moving boat ADCP measurements, or uncertainty estimates from the Interpolated Variance Estimator (IVE) for mid-section measurements) of the measurement (assume excellent < 2 percent; good <5 percent; fair <8 percent; poor = 10 percent; in the case of indirect measurements, use the percent uncertainty assigned by the analyst), un-approve the measurement, revise the discharge value, and re-approve the measurement. Note: The criterion for a revision of a discharge measurement is associated directly with the uncertainty assigned by the hydrographer.
2. Document revisions in SIMS "Revision History" element.
3. Evaluate need for revision of rating or shift (See "Surface Water: Revision due to change in stage-discharge rating or shift" section).
4. Evaluate need for revision of peak-flow file (See "Surface Water: Peak Extremes Revision" section).

Surface Water: Revision due to change in stage-discharge rating or shift

If the associated change in rated discharge over the affected range in stage is larger than twice the maximum uncertainty of applicable discharge measurements (as specified by the hydrographer and assigned in NWIS) used to define that portion of the rating or shift, approved periods of record affected by the rating (or shift) change must be revised.

Scenario 1: Ongoing record computation identifies an error in definition of an approved stage-discharge rating (or shift) curve; the time period affected by the change contains approved data.

Steps to take:

1. Determine the maximum uncertainty of measurements used to define the rating segment or shift curve being revised. Refer to measurement uncertainties assigned in NWIS, as specified by the hydrographer(s) making the measurements (assume excellent <2 percent; good <5 percent; fair <8 percent; poor = 10 percent).
2. Compute the percent change in computed unit discharge as [(Q_new - Q_old)/Q_old x 100] for the applicable scenario below.
1. Revised rating curve: for all gage heights from the lower to upper extreme of the affected range in stage.
2. Revised half-house type shift: for all gage heights from rating merge point (zero shift) to lowest gage height experienced during the approved period.
3. Revised truss-type shift: for the maximum shift value defined by the truss curve.
3. If any resulting percent changes in rated (or shift-adjusted) discharge exceed twice the maximum uncertainty of measurements used to define the rating or shift (number 2 above), un-approve period, reanalyze all aspects of the affected period (see "Surface Water: Reanalysis" section), mark affected computed discharges with "Revised" qualifier [AQUARIUS, document the revised period with a "Note" [AQUARIUS] and re-approve period.

Surface Water: Revision due to change in velocity-velocity and/or stage-area rating for an Index-Velocity Site

If the associated change in rated discharge over the affected range in stage and/or velocity is larger than twice the maximum uncertainty of discharge measurements (as specified by the hydrographer and assigned in NWIS) used to define that portion of the rating or shift, approved periods of record affected by the rating (or shift) change must be revised. Note: The area produced by a stage-area rating and velocity from the velocity-velocity rating are typically not published and are considered operational data that are only revised as part of a revision in discharge.

Scenario 1: Ongoing record computation identifies an error in definition of an approved stage-area and/or velocity-velocity rating (or shift) curve; the time period affected by the change contains approved data.

Steps to take:

1. Determine the maximum uncertainty of measurements used to define the rating segment or shift curve being revised. Refer to measurement uncertainties assigned in NWIS, as specified by the hydrographer(s) making the measurements (assume excellent <2 percent; good <5 percent; fair <8 percent; poor = 10 percent).
2. Compute the percent change in computed unit discharge as [(Q_new - Q_old)/Q_old x 100] for all gage height and velocity combinations from the lower to upper extreme of the affected range over the period applied.
3. If any resulting percent changes in rated (or shift-adjusted) discharge are larger than twice the maximum uncertainty (percent) of discharge measurement(s) used to define the revised rating segment or shift curve, un-approve period, reanalyze all aspects of the affected period (see "Surface Water: Reanalysis" section), mark affected computed discharges with "Revised" qualifier [AQUARIUS], document the revised period with a "Note" [AQUARIUS] and re-approve period.

Surface Water: Peak Extremes [gage height and discharge] Revisions

Scenario 1: Peak discharge value published in peak flow file was found to be in error. This error is unrelated to Scenario 3, a change in the stage-discharge relation, as described below.

Steps to take:

1. Correction to peak flow file required if error in peak discharge is greater than 10 percent. Corrections for errors less than 10 percent can be made at the discretion of the WSC.
2. Document revisions in SIMS "Revision History" element.

Scenario 2: Peak stage value published in peak flow file was found to be in error.

Steps to take:

1. Correction to peak flow file required if error in peak gage height is greater than 0.04 ft. Corrections for errors less than 0.04 ft can be made at the discretion of the WSC.
2. Document revisions in SIMS "Revision History" element.

Scenario 3: Upper portion of stage-discharge rating for a previously published time period was determined to be in error. A new rating has been developed, but below a certain gage height the new rating is the same as the rating used to publish data during a previous time period. One possible reason for such a change includes new measurements obtained under the same channel and floodplain conditions as the period affected by the rating change that is different from the older rating used initially used to determine peak discharge values. Another reason could be that a significant computation error was found in an historic measurement that warranted a revision to the measurement (see Discharge Measurement Revision scenario) and thus the rating.

Steps to take:

1. Correction to peak flow file required if peak discharge computed from new rating is different from previously approved peak discharges by more than 10 percent. Corrections for errors less than 10 percent can be made at the discretion of the WSC.
2. Document revisions to peak flow file in SIMS "Revision History" element.
3. If corrections are made to any discharge values published in the peak flow file, then all time-series discharges must be recomputed using the portion of the new rating that is different from the previously used rating for the entire period believed to be associated with the new rating. Un-approve period, reanalyze all aspects of the affected period (see "Surface Water: Reanalysis" section), mark affected computed discharges with "Revised" qualifier [AQUARIUS], document the revised period with a "Note" [AQUARIUS] and re-approve period. Note: This may lead to some instantaneous peak values in the time-series of discharge to be different, up to 10 percent, from peak discharges in the peak flow file due to decisions made by the WSC (see number 1 above).
4. If a discharge revision indicates that the annual peak discharge occurred during a different storm event than previously published in the peak-flow file, the peak-flow file will be amended with the new-peak date, whether or not the revised peak value exceeds the 10 percent criterion.
5. Document revisions to peak flow file in SIMS "Revision History" element.
6. If no corrections are required to discharge values published in the peak flow file, then it is not recommended to activate the new rating or recompute any of the time series discharges.

Surface Water: Revisions in precipitation data

Scenarios: The following are examples of scenarios that might require a revision to approved precipitation data. Revisions can be made at the discretion of the Water Science Center for changes less than these criteria, if desired. In the case where data have been approved prior to a calibration test supporting that period and the calibration test fails, all data collected after the last successful calibration should simply be deleted from the database; such data is considered to be discredited (no revisions/corrections possible) but the action should be documented in SIMS "Revision History" element. Situations that might result in data eligible for revision of approved precipitation data:

1. An error in processing IVs is found. For example, test tips or calibration amounts are inadvertently transmitted requiring a correction to the running total. The correction was either not made or made incorrectly. Similarly, data that was cumulative may have accidentally been processed as incremental data and was approved before the error was discovered.
2. IV data are added. For example, data transmissions were missed for one or more days and the data had been approved prior to loading the EDL data.
3. An error in processing DV statistics is found. An example of this type of error would be when the processor was incorrectly set up to sum amounts from noon to noon instead of midnight to midnight (IVs are correct but computed DV statistics would be wrong).
4. DV statistics data are added. For example, a clogged funnel may slowly drip over one calendar day. The IVs were considered invalid and deleted and thus no DV statistic was computed. Later, it was determined that the daily total was accurate even though the IVs were not. Similarly, a multi-day gap in the data exists but is bounded by the same cumulative value; in this case the intervening days that were originally shown as missing should be replaced with Daily totals of 0.00 inches.
5. An error in an estimated DV statistic is found. For example, DV statistics were estimated using techniques allowed in OSW 2006.01. A computational error was discovered or data from other gage(s) used to make the original estimate have changed forcing a revision to be considered.

Steps to take: Periodic data has been approved:

1. Unsuccessful calibration: Delete unapproved data collected since previous successful calibration. If there are approved data since previous calibration un-approve period, delete that intervening data, document the revised period with a "Note" [AQUARIUS] and re-approve period.
2. Error in IVs found (scenarios 1&2 above) in excess of 20% and >0.05 inch: Un-approve period, correct totals or merge backup data for the affected period, mark affected values with "Revised" qualifier [AQUARIUS], document the revised period with a "Note" [AQUARIUS and re-approve period.
3. Error in DV statistics found (scenarios 3&4 above) in excess of 10% and >0.05 inch: Un-approve period, re-compute period as needed, mark affected values with "Revised" qualifier [AQUARIUS, document the revised period with a "Note" [AQUARIUS] and re-approve period.
4. Error in estimated DV statistics (scenario 5 above) found in excess of 20% and >0.10 inch: Un-approve period, re-compute period as needed, mark affected values with "Revised" qualifier [AQUARIUS], document the revised period with a "Note" [AQUARIUS] and re-approve period.

Surface Water: Reanalysis

Some of the revision scenarios presented above require a reanalysis of the affected record period. Essentially, if the record period requires a reanalysis, all aspects of the record computation are to be examined, and if necessary, revised in accordance with criteria described above. Below are the general steps that need to be taken during any reanalysis.

1. Reanalyze includes re-analysis of gage height corrections, shifts (revised rating curve); re-computation of all discharge IVs affected by the revision (as defined by the upper and lower gage-height extremes of the rating segment or shift curve being revised); and any estimated IVs or DV statistics within the period of record being re-worked.
2. Reanalyze includes revision of annual peak(s) in the PFF, if a recomputed peak differs from original peak by more than 10% (See "Surface Water: Peak Extremes Revision" section). As noted previously, corrections for errors less than 10% can be made at the discretion of the WSC.
3. If the period of computed record being revised was used as the basis for estimating adjacent periods of missing or ice-affected record (hydrographic trend), reevaluate those adjacent estimated periods and revise as appropriate (to maintain continuity of the record).
4. If the period of computed record being revised was used as the basis for estimating periods of missing or ice-affected record at another streamgage (hydrographic comparison), and revised DV statistics for the index site changed by more than 50 percent, reevaluate estimated periods for the dependent site and revise as necessary (based on updated hydrographic comparison) to maintain relational continuity between the sites.
   Note: The 50 percent threshold (index site) is consistent with Novak (OFR 85-480), and is considered a reasonable threshold to trigger reevaluation of estimated DV statistics at the dependent site.
5. All revisions to time-series data in AQUARIUS are to be mark with "Revised" qualifier [AQUARIUS], and the revised period is to be documented with a "Note" [AQUARIUS].

Surface Water: Example Revision Statements

As outlined in the guidance presented above, revisions to time-series data are to be documented in AQUARIUS using a "Note" in the Data Corrections Toolbox, and revisions to other data types are to be documented in the "Revisions History" element in SIMS. Below are some example statements to be used in either the "Note" or "Revision History" element. These example statements contain four fundamental pieces of information that all revision statements must have:

• What parameter was revised?
• What period was revised?
• When was the revision made?
• Why was the revision made?

The examples below are meant to be used to guide the development of any surface-water data revision statements and all statements must begin with "Revisions:".

Examples:

Revisions: Discharge for the period Nov. 3, 2014, to Dec. 12, 2014, was revised on Dec. 22, 2015, based on changes to the stage-discharge rating.

Revisions: Gage height and discharge for the period Nov. 3, 2014, to Jul. 12, 2015, were revised on Dec. 22, 2015, based on changes to the gage datum.

Revisions: The maximum discharge for water year 2013 was revised on Dec. 22, 2015, based on revision of indirect measurement made following the flood of Aug. 1, 2013.

Water Quality Revisions Guidance

Contents:

Water Quality: Revision due to error in the value of an analytical result or field measurement
Water Quality: Revisions due to error in sample-level body-part and taxa identification data
Water Quality: Revisions of time-series records
Water Quality: Revisions of surrogate regression models
Water Quality: Revision Statement

Water Quality: Revision due to error in the value of an analytical result or field measurement

This guidance applies only to result-level values and to sample-level body part code and taxa identification data. These are the most critical pieces of metadata for interpretation. No specific revision criteria or revision documentation requirements are provided for other result-level or sample-level information. Centers are encouraged to make changes to these data when an error is found that is considered to be significant.

For discrete WQ data, the criteria for revision depends on the analyte and would be 2 times the acceptable uncertainty listed in the project-specific data-quality objectives. In the absence of information on uncertainty, any changes to approved data (revisions) should be documented.

The following scenarios describe situations that could result in a value being revised after approval. These scenarios are by no means all-encompassing but instead serve as examples; there are many other scenarios.

Scenario 1, National Water Quality Laboratory error: Review of a laboratory method (schedule) using laboratory and / or Branch of Quality Systems quality assurance records identifies an error associated with a method for a limited time period. The lab wishes to revise the analytical results from a specific time period using the new information. The previously stored value has already been approved by the Center and the data-quality indicator (DQI) code has been set to "R".

Steps to take:

1. Because the National Water Quality Lab (NWQL) is making the correction, data will be reloaded by NWQL into NWIS-QWData with an accompanying result-level laboratory comment to indicate the date and reason for the revision. The DQI code reverts to "provisional" (DQI code = "S").
2. The new result values need to be re-reviewed and the DQI code set to the appropriate review status by the Center.

Scenario 2, Laboratory error: A contract lab has provided files to revise analytical results that they provided to a Center project. The previously stored value has already been approved by the Center and the data-quality indicator (DQI) code has been set to "R".

Steps to take:

1. Data are reloaded by the Center into NWIS-QWDATA with an accompanying result-level laboratory comment to indicate the date and reason for the revision. The DQI code reverts to "provisional" (DQI code = "S").
2. The new result values need to be re-reviewed and the DQI code set to the appropriate review status by the Center.

Scenario 3, field error: It is discovered that an error was made when a dissolved-oxygen field measurement was entered in NWIS-QWData; the field values were reviewed and approved before the error was discovered.

Steps to take:

1. Correct the field value in QWData.
2. Include an accompanying result-level field comment to indicate the previous value, date, and reason for the revision.
3. Insure that result value receives a DQI code set to reviewed and approved (DQI = "R").

Water Quality: Revisions due to error in sample-level body-part and taxa identification data

These scenarios describe situations that could result in sample-level metadata being revised after approval of water quality results.

Scenario 1, Incorrect sample-level code: A sample containing approved result data is discovered to be assigned an incorrect body part code.

Steps to take:

1. The sample body part code is corrected.
2. A statement describing the date of change, the reason for the change, and the original code setting is added to the sample level comment field.

Scenario 2, Incorrect sample-level code: A sample containing approved result data is discovered to be assigned an incorrect code for taxon identification.

Steps to take:

1. The taxa identification code is corrected.
2. A statement describing the date of change, the reason for the change, and the original code setting is added to the sample level comment field.

Water Quality: Revisions of time-series records

The computation of time-series records for water temperature, pH, dissolved oxygen, specific conductance, and turbidity follow the guidance and procedures presented in Techniques and Methods Report 1-D3 (Wagner and others, 2006). The computation of time-series records of optical nitrate follow the guidance presented in Techniques and Methods 1-D5 (Pellerin and others, 2013). These reports specify the data-correction criterion and the maximum allowable limit (MAL) for a constituent. For time-series records not covered in these reports, the analyst can estimate a data-correction criterion equal to the sensor accuracy and a maximum allowable limit (MAL) equal to the sensor accuracy multiplied by 3.

Scenario 1, new Information: A subsequent field visit provides new information that indicates the applied correction (Techniques and Methods Report 1-D3; Techniques and Methods 1-D5) was incorrect for approved water-quality records. Using the new information, the correction for the period is re-calculated.

Scenario 2, computation error: Periodic record audit identifies an error in the computation of the correction (Techniques and Methods Report 1-D3; Techniques and Methods 1-D5) that was applied to an approved water-quality record. The error is addressed and the correction for the period is re-calculated.

Scenario 3, recording error: Periodic record audit identifies an error in the recorded field notes used to determine the correction for an approved water-quality record. The correction for the period is re-calculated.

Steps to take:

1. Revision is less than two times the data-correction criterion: If the difference between the old and new corrections is less than twice data-correction criterion specified for the parameter, it is up to the WSC to decide if the old correction is removed and the new correction is applied according to USGS guidelines.
2. Revision is more than two times the data-correction criterion but less than the maximum allowable limit: If the difference between the old and new corrections is greater than twice the data-correction criterion specified and the new correction is less than the MAL for the parameter:
1. Unapprove affected period, remove the old correction, and apply the new correction according to USGS guidelines.
2. Mark the affected instantaneous values with the "Revised" qualifier and re-approve the period.
3. Document the revised period in NWIS using a "Note".
3. Correction is greater than the maximum allowable limit: If the difference between the old and new corrections is greater than twice the data-correction criterion specified and the new correction exceeds the MAL:
1. Unapprove affected period, remove the old correction and apply the new correction and delete the record greater than the MAL, in accordance with USGS guidelines.
2. Mark affected period of remaining instantaneous values with "Revised" qualifier and re-approve the period.
3. Document the revised period (includes the revised and deleted values) in NWIS using a "Note."

Water Quality: Revisions of surrogate regression models

The computation of time-series records for sediment concentrations of interest using surrogate regression models are presented in Techniques and Methods Report 3-C4 (Rasmussen and others, 2009) and Techniques and Methods Report 3-C5 (Landers and others, 2016). For time-series records not covered in these reports, the analyst can use the estimated standard error for the model to determine if a revision is necessary. Requirements and guidance for revising surrogate suspended-sediment models is outlined in Attachment A, steps 4-7 of Office of Surface Water Technical Memorandum 2016.07 and Office of Water Quality Technical Memorandum 2016.10.

Scenario 1: A surrogate regression model has been revised and a period of approved computed concentration is significantly affected.

Steps to take:

1. Apply the revised surrogate regression model to recompute the desired time-series record for the time period affected by the model change and flag these unit value data using the "Revised" qualifier in NWIS. Document the revised period with a revision statement "Note" [AQUARIUS] and re-approve period.

Water Quality: Revision Statements

As outlined in the guidance presented above, revisions are to be documented in either AQUARIUS using a "Note" in the Data Corrections Toolbox, or in the specified comment field in QWDATA. Below are some example statements to be used in either the "Note" or the result-level comment field. These example statements contain the fundamental pieces of information that all revision statements must have:

Time-Series data with revision statements in an Aquarius Note.

1. The parameter that was revised.
2. The period that was revised.
3. Date the revision was made.
4. Why the revision was made.

Discrete data with revision statements in a result-level comment.

1. Date the revision was made,
2. The reason the value was revised revision was made,
3. The original value, if possible. This may not be practical for batch reloads.

The examples below are meant to guide the development of any water quality data revision statements and all statements must begin with "Revisions:"

Examples:
Revision to an NWQL value:

Revisions: Result level lab comment for parameter 62065. Value reloaded per NWQL Rapi-Note 14-25

Revision to a field value:

Revisions: Result-level field comment for 00095. Changed value from 779 to 797 on Feb. 23, 2012, to correct a typographical error.

Revision to a time series record:

Revisions: Suspended-sediment data for the period Nov. 3, 2014, to Jul. 12, 2015, were revised on Dec. 22, 2015, based on changes to the regression model.

Revisions: Dissolved oxygen records for the period Nov. 3, 2014, to Jul. 12, 2015, were revised on Dec. 22, 2015, a calibration error was discovered which changed the required data correction.