IIn Reply Refer To:
Mail Stops 412, 411, and 415

March 19, 2012


Subject: Policy for release of continuous water-temperature data recorded from secondary sensors at streamgages and wells


This memo establishes policy and guidelines for quality assurance (QA) and review of water- temperature data from secondary thermistors. (A secondary thermistor is a temperature sensor that has a primary purpose to provide temperature correction data to another sensor. A primary thermistor is a sensor designed to measure water temperature.) This policy applies to the continuous secondary temperature records displayed on the Internet, provided to cooperators, or archived in the National Water Information System (NWIS) database.


Some Water Science Centers (Centers) have posted real-time continuous water-temperature data on NWISWeb without review until the time limit for provisional data expires, after which these temperature data disappear from NWISWeb. Centers justify this practice by noting that the data are from secondary thermistors built into submersible pressure transducers or other instruments, so the data provide a service to customers and cooperators while being available at virtually no extra cost. It is inconsistent for unapproved data to appear on a U.S. Geological Survey (USGS) web-data portal with no intent to quality assure the results, because our partners and the public have a right to expect all data to be of similar reliability. The inconsistency remains even if the data are marked provisional and do not remain permanently accessible, as allowing this practice would degrade the meaning of "provisional" as we apply it to other data series (WRD 95.19/WRD 99.34).

Some cooperators request these data even though the secondary thermistors are potentially less accurate than standard sensors and have not been subjected to standard quality-control procedures; such records do not receive standard USGS QA and review and approval. The current practice of posting real-time continuous water temperature data from secondary sensors on NWISWeb as "operational instantaneous data" (i.e., temporarily published provisional data) without review does not meet USGS Fundamental Science Practices for information product review (SM 502.4) nor Office of Water Quality policies for quality assurance and control, and data review, approval, archival, and release to the public.


The display of real-time temperature data on NWISWeb for a limited time without review should be discontinued immediately. Centers should choose from the following alternatives for managing these data;

  1. Hold the data for internal use only and discontinue public display to NWISWeb.

  2. Process the data using the same QA as for any other continuous data series, thus allowing the data to be displayed on NWISWeb.
    The Center is encouraged to make the decision to provide quality-assured secondary-sensor continuous temperature data for all sites wherever the stage or groundwater level equipment allows and the secondary sensor meets accuracy requirements, so that the cost would be included in the base rate for all gaging stations or groundwater wells. The Water Mission Area envisions continuous temperature data at streamgages and wells becoming as common and as useful as real-time communication of other continuous water-quality data. Climate variability and change, for example, is one of six science directions in the USGS Science Strategy (USGS, 2007) and accurate water temperature data is an important contribution to the science.


Standard procedures for temperature measurement with continuous water-quality monitors generally produce "good" quality data, with an overall uncertainty of less than 0.5°C using single-purpose thermistors with calibrated accuracy within ±0.2 °C (Wagner and others, 2006; Wilde and others, 2006). Many submersible pressure transducers, acoustic velocity meters, and other sensors used in water-resource monitoring and research include a thermistor that allows the instrument to maintain a stable calibration through a wide range of temperatures. However, such secondary thermistors may be accurate to 0.5 - 1.0 °C because an accurate result for the primary target variables may not require the same temperature accuracy as would be expected from a single-purpose temperature sensor.

Our scientists routinely program field instruments to transmit temperature and any other secondary sensor observations, to NWIS to be used as resources for assuring the quality of the stage, velocity, groundwater level, and other primary data. Once these secondary temperature data are in NWIS, they become easy to display on NWISWeb as provisional, and potentially misleading, record.

For secondary temperature records, the issues of concern are (a) understanding what the data represent, (b) the quality of the data (c) data storage and archival, (d) clear reporting to data users (Dunham and others, 2005; Wagner and others, 2006). For stream applications, pressure transducers are installed at locations where the water level represents the flow approaching some identified hydraulic control but not necessarily where the flow is well mixed. A pressure transducer located in a stilling well, in backwater, or in direct sunlight at certain times of day could provide a representative stage record but a temperature record that does not represent the thermal state of the stream cross section. Likewise, a pressure transducer located adjacent to a fracture or in a shallow well with low hydraulic conductivity may not represent the thermal state of the target aquifer. Several factors could cause the recorded temperature to lag behind changes in stream water temperature, which might cause the daily temperature range or maximum to be underestimated (Dunham and others, 2005).

Thermistors are characteristically stable and accurate for long time intervals, but eventually drift. One manufacturer indicates the most common cause of drift is a gradual change in the crystal structure of a thermistor exposed to moisture, which causes the electrical resistance to decrease and the indicated temperature to increase (Quality Thermistor, Inc, 2007). Experience among USGS Centers has shown measurable differences between separate thermistors installed at the same site. In addition to such bias, failing thermistors are known to start giving erratic readings or to abruptly shift out of range for environmental water.

Routine quality checks for stream stage, stream velocity, or groundwater level compare the indicated values against measurements from an independent instrument. If similarly independent temperature measurements are overlooked for secondary thermistors, the resulting data quality is unknown.


For secondary temperature records, the following guidelines ensure (a) that data represent the location as designated by the site description, (b) the quality of the data meets minimum USGS requirements, (c) data are stored and archived according USGS Fundamental Science Practices, and (d) data are clearly reported to data users. The Centers have two valid alternatives for recording temperature data in NWIS and displaying records of water temperature from secondary thermistors to NWISWeb:

  1. Treat data from secondary thermistors as ancillary information for internal use only. Set the NWISWeb access level for the relevant data descriptors (DD) to 2 (all USGS) or 3 (host Center only), thereby blocking the data from reaching the public. Though these data will remain in the NWIS database, they should not be in a DD marked "primary" and should never be flagged as approved. Temperature-specific field tasks are not necessarily needed, but the station analysis or other Center records should document what these data represent.

  2. Process and release the temperature data using standard USGS procedures for continuous water-quality monitors (Wagner and others, 2006; and as cited below from Wilde and others, 2006). For deployment in surface water and groundwater:

    1. The accuracy of the secondary thermistor must be within ±0.2°C, as required for normal field instruments (Wilde and others, 2006; USGS Office of Surface Water Technical Memorandum 2010.07) and calibration of the secondary thermistor must be checked periodically, as appropriate, according to guidance in Wilde and others (2006), Wagner and others (2006), Gibs and others (2007), and USGS Office of Surface Water Technical Memorandum 2010.07).

    2. The secondary sensor should be such that the temperature data demonstrably represent the section or point of the hydrologic environment intended to meet project and data-collection objectives. Included, for example, are continuous point measurements in surface-water bodies or at targeted depth intervals in wells.

As they pertain to continuous water-quality measurements, calibration-check procedures developed to ensure that an accuracy of±0.2°C is monitored and maintained over the duration of thermistor deployment are given in Wilde and others (2006, Section 6.1.2.B, p. 7) as follows: "For temperature measurements, 'calibration' [as defined in footnote 2] refers to a comparison or accuracy check at specified temperatures against a thermometer that is certified by the National Institute of Standards and Technology (NIST), or is manufacturer-certified as NIST traceable. Calibration [checks] should be performed... every 6 to 12 months, depending on the manufacturer's recommendation... In the case of continuous monitors, a nonmercury calibration thermometer can be used in the field to check or monitor temperature readings whenever other field-measurement sensors are calibrated. See Wagner and others (2006) for spe­cific guidelines for continuous monitors..." This guidance further explains (p. 10) that the time interval between calibration checks depends also on the environmental conditions to which the thermistor is exposed, including the amount of use and abuse to which it has been subjected. Gibs and others (2007) recommends an accuracy check of the temperature thermistor every three months and states that if the difference between the readings does not fall within the manufacturer-specified accuracy, the instrument should be returned to the manufacturer for repair or replacement. Wagner and others (2006, pp. 31-33) provide specific information for making corrections for calibration drift and sensor fouling, and recommend periodic cleaning of thermistors and other sensors for any instrument that is deployed for continuous monitoring.

Although Wagner and others (2006) do not address groundwater applications directly, the criteria and general processes for calibration, data correction, and ratings accuracy, may be applied to, or adapted for, groundwater continuous water-quality data. Wagner and others (2006, p. 2), moreover, provide guidelines for locating continuous monitor sensors to represent stream, lake, and estuarine environments. Wilde and others (2006, Section 6.1.3.C) and Gibs and others (2007, Section 6.8.3.B) describe methods for making temperature measurements that are representative of the groundwater environments intended for study, and that assume good aquifer communication with the well has been verified.

Questions about this policy should be directed either to Franceska Wilde or Stan Skrobialowski, Office of Water Quality.

Donna N. Myers /s/
Chief, Office of Water Quality

William L. Cunningham /s/
Acting Chief, Office of Groundwater

Robert R. Mason /s/
Acting Chief, Office of Surface Water

Distribution: All WRD Employees


Dunham, Jason, and others, 2005, Measuring stream temperature with digital data loggers, a user’s guide: USDA Forest Service General Technical Report RMRS-GTR-150WWW, 15 p., http://www.treesearch.fs.fed.us/pubs/9476

Gibs, Jacob, Wilde, F.D., and Heckathorn, H.A., 2007, Use of multiparameter instruments for routine field measurements (ver. 1.1): U.S. Geological Survey Techniques of Water-Resources Investigations, book 9, chap. A6, section 6.8, accessed__March 7, 2012,__ from http://pubs.water.usgs.gov/twri9A/.

Quality Thermistor, Inc, 2007, NTC [Negative Temperature Coefficient] thermistor design guide, Summer 2007: http://www.hy-line.de/fileadmin/hy-line/sensorik/hersteller/datasheets/qti_ntc_thermistor_design_guide.pdf, accessed September 20, 2010.

U.S. Geological Survey, 2007, Facing tomorrow’s challenges—U.S. Geological Survey science in the decade 2007–2017: U.S. Geological Survey Circular 1309, x + 70 p.

Wagner, R.J., and others, 2006, Guidelines and standard procedures for continuous water-quality monitors—Station operation, record computation, and data reporting: U.S. Geological Survey Techniques and Methods 1—D3, 51 p.

Wilde, F.D., 2006, Temperature (ver. 2.0) in National field manual for the collection of water-quality data: U.S. Geological Survey Techniques of Water-Resources Investigations, book 9, chap A6, sec. 6.1, March 2006, accessed March 9, 2012, from http://pubs.water.usgs.gov/twri9A6/.

U.S. Geological Survey Manual 500.24 - 502.4 - Fundamental Science Practices: Review, Approval, and Release of Information Products, accessed February 9, 2011 at http://www.usgs.gov/usgs-manual/500/502-4.html


OSW Tech Memo 2010.07 Independent Water Temperature Measurement for Hydroacoustic Measurements

OSW Tech Memo 2006.01 (revised 12/2009) Collection, Quality Assurance, and Presentation of Precipitation Data

WRD Policy memo 99.34 Quality Assurance Measures for Serving Real-time Water Data on the World Wide Web
Memorandum 95.19 established a policy that the nature of provisional data available on WWW would be clearly identified (this memo is referenced in 99.34 but is not available on the memorandum web site)