Ultraviolet Radiation – 07 METHODS Measuring and modeling solar radiation and spectral diffuse attenuation

Ultraviolet Radiation and Bio-optics in Crater Lake, Oregon, 2005


Measuring and modeling solar radiation and spectral diffuse attenuation

Published measurements of Crater Lake Kd,380 from the 1960s were obtained from several sources (Smith and Tyler, 1967; Tyler & Smith, 1970; Smith et al., 1973). A LI-COR scanning radiometer (model LI-1800uw) was the primary instrument for the new measurements of UVR irradiance and attenuation reported here (performance reviewed in Kirk et al., 1994). The self-contained programmable scanning radiometer records downwelling cosine irradiance at 2 nm intervals from 300–800 nm with a bandwidth of 8 nm. LI-COR post-collection software provided immersion corrections to maintain accuracy both above and below the water surface. To create a depth-series of spectra the instrument was programmed to scan every two minutes and was then lowered on the sunny side of a small vessel near mid-day to specific depths and held for timed intervals. On many occasions an incident PAR signal was recorded on deck during the underwater scans in order to detect changing sky conditions. For each depth-series of spectra the data were evaluated at specific wavelengths to compute spectral Kd for each pair of depths (typically at 5 or 10 m vertical spacing). The data were carefully examined for anomalies (identified as outliers in both spectral and depth plots of Kd,?) caused by surface waves or clouds; the occasional anomalies were either eliminated by interpolation between adjacent wavelengths or depths. The depth assigned to each Kd was the average for the upper and lower pair of Ed measurements. The LI-1800uw was factory calibrated on 26 May 1995, 18 July 2000, and 23 January 2002.

On 20 August 2001 two other UV radiometers (from Biospherical Instruments) were also used to record water depth and temperature and up to 20 fixed wavelengths of downwelling irradiance (Ed) and upwelling radiance (Lu) using filter-based diode sensors at a rate of 5 spectra per second. A PUV-2500 profiling UV radiometer recorded seven downwelling channels (305, 313, 320, 340, 380, 395 nm with nominal 8 nm bandwidth and PAR, 400–700 nm) plus upwelling radiance in the chlorophyll-a natural fluorescence waveband (center =683 nm). A PRR-800 profiling reflectance radiometer recorded 19 channel pairs of upwelling radiance and downwelling irradiance (340–710 nm) and PAR. The PRR-800 was lowered in its normal orientation to measure Ed and Lu and also lowered inverted to record upwelling irradiance Eu. The data from the Biospherical instruments were binned at 2 m depth intervals using Loge averages. To detect internal radiation sources that could interfere with interpretation of diffuse attenuation measurements we measured spectral reflectance ratios for a range of depths using the PRR-800 radiometer. To reduce near-surface noise cause by waves and ripples we used both running averages of Kd (combining several adjacent depths) and polynomial regression of Ln(Ed) versus depth from which Kd was then calculated.

On this date we also computed incident irradiance using a radiative transfer model (RTBASIC, Biospherical Instruments, Inc.; see Madronich, 1993 and Biospherical Instruments, 1998, for more details) for comparison with the LI-1800uw incident spectra. The parameters for the 8-stream disort model were set to account for noon PDT conditions near the time that optical profiles were collected: current solar zenith angle (34.4o), barometric pressure (774 mbar based on elevation above sea level), nominal albedo (5%) and current column ozone (313 DU, URL: http://toms.gsfc.nasa.gov/). To compute a spectrum with a bandwidth comparable to the LI-1800uw (8 nm) and comparable reporting interval (2 nm), model values with 1 nm bandwidth were first generated at 2 nm intervals from 296–804 nm. These initial model values were then converted to represent an 8 nm bandwidth by calculating geometric means spanning 8 nm around each 2 nm interval.

To calculate PAR (400–700 nm) from LI-1800uw data, values reported as W nm-1 m-2 were summed and then multiplied by 2 nm per record to get the energy in the band, W m-2. Quantum PAR irradiance (?Mol m-2 s-1) at the surface was calculated from W m-2 by multiplying by 4.60 (Kirk 1994b). To interpolate to an intermediate wavelength (e.g. 305 nm from 304 and 306 nm measurements) we used the geometric mean of measured irradiance. To reduce surface noise caused by waves and ripples the scans near the surface were repeated and averaged, or computed using the incident scan as the upper value after reducing it by 5% for nominal reflectance.

<< previousnext >>