¶ Normal Incidence Multifilter Radiometer (NIMFR)
The Normal Incidence Multi-Filter Radiometer (NIMFR) is a ground-based instrument that provides a time series of the shortwave spectral direct normal irradiance. Additionally, there is a broadband silicon detector that can crudely measure the direct normal broadband irradiance. The direct normal irradiance readings are measured at 6 different wavelengths: 415, 500, 615, 673, 870, and 940 nm (nominal width of 10 nm) and has a field of view of 5.7o. The output from the NIMFR should be VERY similar to that of the MFRSR. Direct comparison of the MFRSR with the NIMFR is suggested (Compare NIMFR E13 with MFRSR E13 and MFRSR C1 direct_normal_narrowband_filter#). The NIMFR is found at the SGP Central Facility and NSA Barrow.
¶ Instrument Design
The instrument consists of a “head” from an MFRSR to which a collimating tube is attached. This assembly is mounted in a solar tracker so that the instrument points directly at the sun (i.e., the collimating tube points at the sun). The collimating tube has a narrow field of view designed to admit only the direct normal irradiance; this radiation hits the detectors housed in the MFRSR head. A very small amount of diffuse radiation is also admitted by the tube and sensed by the detectors. The amount of diffuse radiation that strikes the detector is so small in relation to the direct beam radiation that it is usually neglected. For the spectral measurements, the detector for each wavelength channel consists of a combination interference filter/silicon photodiode (responsive to radiation between 300 and 1100 nm). The instrument’s head is kept at a relatively constant temperature of about 35oC.
The NIMFR is mounted on a solar tracker that is usually about 2 or 3 meters above the ground. At a specified time interval (20 sec at the NSA sites), the instrument takes a reading (six spectral irradiances and one broadband irradiance). The readings are stored on a nearby data logger.
¶ Uncertainty
The uncertainties in the measurement depend on the state of calibration of the instrument. The NIMFR instrument is calibrated using a standard lamp that is uncertain to +/- 5%. With proper calibration, the broadband measurement is accurate to about 5%. The calibration stability of the instrument depends on the several factors, including the stability of the interference filters used to make the spectral observations and the temperature stability of the MFRSR head. Although there have been problems in the past with interference filter stability (the transmission of these devices can change over time), these problems have been largely solved for the NIMFRs ... or so we hope. Time will tell if the problem has indeed been solved, but so far the filters are staying stable.
The head temperature affects the magnitude of the signal coming out of the MFRSR head, and this sensitivity has been estimated to be a 0% to 2% change in output per 1oC change in head temperature. It is not unusual for the head temperature to fluctuate over a range of +/- 1oC over the course of a day, and +/- 3oC over the course of a year. Such fluctuations can cause the calibration to change significantly in step with the temperature variations. To date, this problem has not been solved. The extent to which this problem affects measurements from any given NIMFR has not been determined. All we can say is that the error could be as large as +/- 6% for some wavelengths. Not all instruments will exhibit errors of this magnitude and for a given instrument the error may be small for one wavelength channel and large for another. The 6% figure cited above should be considered a worse-case scenario.
¶ Calibration
The NSA C1 NIMFR is removed once a year every year and sent in for calibrations around the time of year when Barrow goes into 24 hour darkness. The instrument is inspected for problems and calibrated, then reinstalled when daylight returns in the spring.
¶ Known Data Quality Problems
Data quality problems are the same with the NIMFR as they are with the MFR and MFRSR. Specific examples from past problems can be found below:
MFRSR vs. NIMFR
Comparisons with NIMFR at high latitudes may show a problem around local solar midnight. This is because the MFRSR will stop tracking and recording data +/- 2 hours of local solar midnight. Please see Past Problems for MFRSR.
Cleaning NIMFR
The SGP CF NIMFR has increased its daily lense cleaning from once to twice per day (starting 9/2006). This was implemented because a study discovered a jump in the signal following a window cleaning.
¶ Examples
NSA has some interesting working conditions, and often can create interesting conditions. Here are some examples NSA Site Examples:
Lead ice creating ice fog
05/16/2007 was a nice clear day and night in Barrow.
Towards the north, where the sun doesn't quite set these days, is the ocean. These days, there is often a lead (crack in the sea ice) near the shore (Figure 1), and ice fog comes off that lead. The ice fog is nicely visible in the NIP data from the previous day (Figure 2), as well as in the camera images shown below (Figure 3).
