Advances in measurement of shortwave and longwave radiation at sea

About two decades ago an ambitious international air-sea interaction experiment took place in the western Pacific ocean, the TOGA-COARE (Coupled Ocean-Atmospheric Response Experiment). Following the lead of WOCE, a goal was set for the accuracy required for measurements of the net energy exchange between ocean and atmosphere, of 10 Wm-2 over weekly to monthly time-scales. The main consideration was the extreme sensitivity of coupled ocean-atmosphere models to the energy fluxes which at that time required arbitrary “flux adjustments” to avoid unrealistic excursions of the physical variables (e.g. SST). Following the experiment, an international air-sea interface working group (or “flux group”) reviewed the accuracy of observations from the 13 ships, 5 aircraft, and the IMET mooring, of the various components of this energy exchange. Suffice to say that there were some surprises, documented in flux group reports and summarized in Weller et al. (2004).

The exercise led to substantial improvement over time in both instrumentation and methodology (Bradley and Fairall, 2007). The most intransigent were the two radiative components. The net longwave exchange is the balance between the emitted radiation, and the downwelling component emitted from clouds and atmospheric constituents, particularly water vapour in the tropics. The former is calculated from SST using the Stefan-Boltzman law, while the latter is measured with an infra-red radiometer (a pyrgeometer) whose principle of operation not well understood in 1992. Disagreements between individual instruments ranged from 12 Wm-2 during the night to 50 Wm-2 during the day, well outside our target accuracy for all flux components combined. Several sources of error were identified, and the fundamental theory of the pyrgeometer re-examined (Fairall et al. 1998). Based on what we now know about possible faults in pyrgeometer construction and calibration, Colbo and Weller (2009) estimate the fundamental uncertainty of the instrument to be about 4 Wm-2 for daily averages. During a recent intercomparison at sea, three pyrgeometers showed instantaneous differences of 5 Wm-2 during the day and 3 Wm-2 at night.

Shortwave (solar) radiation is measured with the pyranometer, an apparently simple instrument with a single linear output signal, which has existed in its present form for around 70 years. Modern instruments have a specified accuracy of about 2%, which nevertheless amounts to over 20 Wm-2 at the peak of a clear day. Side-by-side intercomparisons of several COARE instruments revealed that individual pyranometers differed by up to 12% or over 100 Wm-2 at midday. More recent uncertainties in calibration methods led to an extensive study of the factors which can cause disagreement between instruments. These are now largely resolved; during a 2-hour clear period in the above intercomparison cruise, when the average radiance was about 1040 Wm-2, 3 pyranometers from different calibration facilities were found to agree within 5 Wm-2.