The microwave radiometer measures the emitted energy from the earth/atmosphere system in the microwave wavelength region (1-100 GHz). Since the Rayleigh-Jeans approximation holds in the microwave wavelength regime, the emitted energy is proportional to the temperature of the radiator to a first order; therefore, intensity is synonymous with temperature at these wavelengths.

Rayleigh-Jeans equation:

E=2kBTcv2

Where:

kB = Boltzmann's constant
c = speed of light
v = wave number

The Nimbus-5 flew in a circular sun-synchronous orbit at 1112 km (600 nautical miles), had a local noon (ascending) and midnight (descending) equator crossing, and an 81° retrograde inclination. Successive orbits crossed the equator at 27° longitude separation. The orbital period was about 107 minutes. Nimbus-5 used an attitude control system which stabilized the spacecraft with respect to the earth and orbital plane, such that the yaw axis pointed normal to the earth and the roll axis aligned with the spacecraft velocity vector, and which also maintained the solar paddles' orientation to the sun. The system permitted fine control of ± 1° in pitch and ± 0.5° in roll and yaw.

ESMR consisted of four major components:

• A phased array microwave antenna consisting of 103 waveguide elements each having its associated electrical phase shifter. The aperture area is 83.3 cm X 85.5 cm. The polarization is linear, parallel to the spacecraft velocity vector.
• A beam steering computer that determines the coil current for each of the phase shifters for each beam position.
• A microwave receiver with a center frequency of 19.35 GHz and an IF bandpass of from 5 to 125 MHz; thus it is sensitive to radiation from 19.225 to 19.475 GHz, except for a 10 MHz gap in the center of the band.
• Timing, control and power circuits.

Principles of Operation

Unlike conical scan instruments such as the SMMR and SSM/I, the ESMR was a cross-scan instrument, with a resolution of approximately 30 km, which measured primarily the intensity of electromagnetic radiation thermally emitted from the Earth's surface at a wavelength of 1.55 cm (19.35 GHz). The instrument recorded radiation from 78 scan positions, and all observations were first converted to equivalent nadir observations.

Sensor/Instrument Measurement Geometry

The Nimbus-5 ESMR recorded radiation from 78 scan positions varying ± 50° from the satellite track every four seconds (Wilheit 1972). The beam width is 1.4° x 1.4° near nadir and degrades to 2.2° crosstrack x 1.4° downtrack at the 50° extremes. For a nominal orbit of 1100 km altitude, the resolution is 25 km x 25 km near nadir, degrading to 160 km crosstrack x 45 km downtrack at the ends of the scan. Full coverage of the entire polar area could be obtained from a sequence of six satellite orbits, or one-half day of good data, if all 78 beam positions were used; however, because of the large disparity in the radiometer field of view from the outer beam position to the middle beam position (70 km x 140 km compared with 25 km x 25 km), only the middle 52 beam positions were used for a swath-angle coverage of ± 30.5° and a minimum resolution of 29 km by 42 km. This swath angle corresponds to a spatial coverage of about 1280 km on the Earth's surface.

Calibration

The radiometer was originally calibrated using hot and cold reference sources. A sky horn measuring the 3 K cosmic background provided the cold-load temperature reference (Tc). The hot-load temperature was provided by reference to a floating ambient termination in the spacecraft. Calibration parameters are gathered from eight scans of data. Calibration temperatures (T_C and T_H) were calculated from multiplex data, and values of four ambient and four cold calibration voltages averaged through the set of eight scans. For each beam position the brightness temperature (Tin) corresponding to voltage (V) was then calculated by:

Additional Calibration

A new calibration procedure was implemented in 1999 for the Nimbus-5 ESMR data at the same time the data were regridded from the orbital radiance data tapes to the polar stereographic grid. The new procedure included the removal of instrument drift and sensitivity jumps.

Studies of temporal brightness temperature variations in the Southern Ocean revealed some unexpected shifts. Because ocean brightness temperatures are expected to have minimal or no seasonal dependence, it was concluded that these could be caused only by calibration or instrument problems. Although brightness temperatures are affected by surface roughness, foam cover, water vapor, and rainfall, none of these have seasonal characteristics. Comiso and Zwally (1980) discuss the procedures used to normalize the data in response to these anomalous shifts.

While regridding orbital data to the polar stereographic grid, unusually high radiances resulting from instrument malfunction were omitted by gridding only values 0 < Tb < 310 K in the Arctic and 0 < Tb < 282 K in the Antarctic. Only the center 52 sets of values (center beam positions) were used, and 13 sets of values at the beginning and end of each data record were ignored; thus, no data are available poleward of approximately 86° latitude.

Instrumental drifts and meanders were adjusted by considering individual histograms of the radiances for each day. The typical histogram has two peaks in its distribution, one for open water and continental ice sheets and another for sea ice and land. The working premise is that the low-radiance wing of the ocean/ice sheet peak and the high-radiance wing of the sea ice/land peak are invariants except for seasonal changes. The seasonal swing on the high radiance end is from about 255 K - 285 K. At the low radiance end, it is about 120 K - 130 K. Multiple linear regression of all data was used to establish a nominal value of the presumed invariant radiance spread between these two wings. All single-day radiances were then adjusted by an automated procedure so that their histogram patterns match the nominal one in its spread.

North and south polar data were treated separately. The automated procedure entailed ascertaining the offset, slope and first five harmonics (ten trigonometric terms) of the annual cycle in the time series of fixed points (with outliers removed) in the threshold of the water/ice sheet distribution and the tail of the sea ice/land distribution. Models of the low and high radiance time series were constructed using the offset and ten trigonometric terms but not the slope, which was deemed to be an instrument drift. After applying the equations below, the remaining instrument drift was less than 0.1 K over the four-year period.

All radiances in each daily grid were then adjusted with the linear transformation:

Tbadjusted = a + b * Tboriginal

Where:

b = (Tbmodelmax - Tbmodelmin)/(Tbdatamax - Tbdatamin)
a = Tbmodelmin - (b * Tbdatamin)

Tbmodelmax is the daily value of the modeled high radiance wing point, and Tbmodelmin is the daily value of the modeled low radiance wing point. Tbdatamax is the daily value of the actual data high radiance wing point, and Tbdatamin is the daily value of the actual data low radiance wing point. When the actual data are near the model values, there is very little correction. The daily adjusted data were then visually inspected and any remaining unsuitable data were discarded. Unsuitable data were regarded as radiances that did not fit the above model.

Further corrections to the calibration were necessary to account for the antenna ohmic loss, which is a function of beam position and the temperature of the phase shifters, and for the effects of side lobes and the different viewing angles. A set of correction parameters for each beam position was empirically determined using ocean data and used to determine the final calibrated brightness temperatures.

Figures 3 and 4 are images of the ESMR brightness temperatures before and after the data adjustment.

The original, uncorrected data files are available by special request from NSIDC User Services.