Estimating UV Trends Satellites

Global estimates of surface UV irradiance Fx as a function of latitude, longitude, and wavelength A have been calculated from satellite measurements of atmospheric backscattered UV radiances and the small amount reflected from the surface. The long-term precision and stability of a satellite instrument's in-flight calibration, especially the single-channel radiances used to estimate cloud transmission and reflectivity, make it very useful for estimating trends in F^. In the absence of a widely distributed, closely-spaced surface network of well-calibrated UV spectrometers, satellite UV irradiance estimates are extremely useful, especially over ocean areas where there are no other measurements.

There are two ways of estimating the UV irradiance reaching the ground from satellite remote sensing of ozone, aerosol, and reflectivity. First, one can enter these quantities into a detailed radiative transfer model to compute cloud transmission CT using Mie theory to approximate the cloud and aerosol properties, in addition to Rayleigh scattering and ozone absorption (Krotkov et al., 1998; 2001). The second approximate method, is to estimate the irradiance reaching the ground for a Rayleigh scattering and ozone absorbing atmosphere FCLEAR, and then add the cloud and aerosol transmission as a correction factor based on the measured fractional scene R (0 < R < 1) and surface reflectivity RG, T ~ (1 - R)/(1 - RG), where 0 < T < 1. The irradiance at the surface is then approximately

The two methods quite closely agree (Krotkov et al., 2001), except when there is enough multiple scattering within a cloud to give enhanced ozone absorption at wavelengths less than about 310 nm where CT is the better estimate. Irradiances from both the CT and the simplified methods are frequently higher than measured irradiance values on the ground, usually caused by an underestimate in the satellite calculation of aerosol amounts and aerosol absorption (Krotkov et al., 1998; 2001; Herman et al., 1999; Kalliskota, 2000). The differences become much less when the aerosol amount is small or is known from ground-based measurements. Other sources of differences between ground-based measurements and satellite estimates of UV irradiance arise from the large satellite field of view (50 x 50 km2 at nadir for TOMS and 13 x 24 km for OMI) compared to the smaller ground-based field of view, in addition to the terrain height differences within a satellite field of view.

A recent comparison of measured UV erythemal irradiance from ground-based measurements and OMI satellite estimates has been made (Tanskanen et al., 2007). The comparison shows that for flat, snow-free regions with modest loadings of absorbing aerosols or trace gases, the OMI-derived daily erythemal doses have a median overestimation of 0% to 10%, and that 60% to 80% of the satellite estimated erythemal doses are within ± 20% compared to ground-based measurements.

Similar errors occur when interpolating between widely separated ground-based stations, where the aerosol, ozone, and cloud amounts vary between the stations. Given the need for global coverage of Fx, and the sparsely located ground-based stations, calculations of Fx from satellite-observed column ozone abundances and cloud reflectivities, which are validated by ground-based measurements, are a useful method for estimating regional, zonal averages, and global UV irradiance trends.

The year-to-year shifts in cyclic weather patterns (e.g., clouds, ozone transport, etc.) by even a tenth of a degree in latitude and longitude (~10 km), strongly affect ground-based UV measurements and their estimates of UV irradiance trends, but have a minimal effect on area-averaged satellite ozone and reflectivity measurements (and the UV estimates derived from them). Therefore, the surface UV changes deduced from ozone amounts and reflectivity measured by satellites, Fx, are expected to be equivalent to those from cloud-filtered, ground-based observations of UV irradiance, and are superior for estimating regional and global changes. Satellite measurements provide both regional and global long-term coverage, which can be used to construct zonal and regional averages and long-term trends, which have much less geophysical variance from clouds than corresponding ground-based measurements. However, most satellite measurements are from low earth polar orbits that pass over a given site only once a day, and so represent the precise cloud conditions for that local time (usually from 10:30 to 13:30 hours), which misses the morning to afternoon variation. For the purposes of UV trends, the estimations are usually calculated for solar noon geometry using ozone amount and cloud reflectivity from sun-synchronous satellite near-noon measurements.

The use of satellite estimates presupposes ground-based measurements for validation and as a bridge between successive satellite instruments, if there are gaps. However, the determination of local UV irradiance is best achieved by ground-based measurements of either the irradiance or the atmospheric properties (ozone, aerosols, and clouds) above the observation site from which the irradiance can be calculated.

Satellite-observed long-term changes in mid-latitude zonal average ozone amounts suggest that there were significant UV increases for both erythemal irradiance and individual UV-B Fx. The zonal average irradiance increase, relative to 1980, for the latitude band between 30°N and 40 °N peaked in 1993 at about 20% (erythemal irradiance) and 40% (305 nm irradiance). Fortunately, these increased percentage changes occurred during the winter months when the SZAs are large, so that the absolute irradiances were comparatively small and the biological effects were minimal. The calculated annual average irradiance increase during 1993 was about 7% and 14%, for erythemal and 305 nm irradiances, respectively. By 2007, ozone had partially recovered so that the irradiance increase moderated to 4% and 8%, respectively. Model calculations show that the recovery is a direct consequence of the implementation of the Montreal Protocol, and its subsequent amendments, limiting the introduction of ozone destroying substances into the atmosphere.

The long-term (30-year) monthly and zonal average (5° bands) ozone time series can be used to estimate changes in monochromatic irradiance reaching the earth's surface using the RAF from Eqs. (5.1) and (5.2), and the estimate of the SZA as a function of latitude and season from Eqs. (5.12) and (5.13). Sample results are shown in Fig. 5.9 for latitude bands centered on 32.5°S and 32.5°N where the ozone change dO3/O3 is approximately -3% over 30 years. Since there has been no significant change in zonal average cloud cover (Herman et al., 2009b) at these latitudes, the increase in 305 nm irradiance dF/F is about 6% over 30 years for the summer months centered on June (Herman, 2009). While this increase is significant, it appears that the decrease in ozone and increase in irradiance has leveled off starting in the late 1990s.

Figure 5.9 Fractional change in 305 nm irradiance dF/F (upper panels) caused by a change in ozone (lower panels) in two latitude bands centered on 32.5°S and 32.5 °N of 5° width. The red dots represent the summer solstice months of December ( - 32.5°S) and June (32.5°N). The monthly average ozone values are available from the GSFC website based on merged data from multiple satellites (

Figure 5.9 Fractional change in 305 nm irradiance dF/F (upper panels) caused by a change in ozone (lower panels) in two latitude bands centered on 32.5°S and 32.5 °N of 5° width. The red dots represent the summer solstice months of December ( - 32.5°S) and June (32.5°N). The monthly average ozone values are available from the GSFC website based on merged data from multiple satellites (

The two principal causes for change in irradiance at the earth's surface, dO3IO3 and dRIR are shown in Figs. 5.10 and 5.11. The change in reflectivity is based on the preliminary analysis of reflectivity (Herman et al., 2009b; Herman, 2009) for the entire 1979 to 2008 period. The current best-calibrated nadir-view zonal averaged LER values are averaged values from the temporally overlapping satellites listed in Table 5.3. The final numbers are expected to be slightly different than those in Fig. 5.11, so that the cloud-transmission correction to the 305 nm irradiance caused by variability in cloud cover will also change in the final analysis.

Carrying Capacity
Figure 5.10 The 30-year percent change in zonal average annual ozone amount as a function of latitude
Figure 5.11 The 30-year percent change in zonal average annual reflectivity caused by clouds and aerosols (preliminary estimate)

Based on these numbers, an analysis for the annual mean change of 305 nm irradiance can be carried out for multiple latitudes with the results shown in Fig. 5.12. The changes in 305 nm irradiance are quite large at higher latitudes in both hemispheres, but especially in the Southern Hemisphere where it amounts to an

Table 5.3 Satellite instruments for ozone and reflectivity

1979 - 1992 Nimbus-7/TOMS (N7) Full global coverage every day. Only a few missing days from1980 to 1992. Ozone plus reflectivity. Near noon orbit

1985 - 2008 SBUV-2 Series (N-9, N-11, Nadir viewing only. Only a few missing days. Ozone N-16, N-17, N-18) plus reflectivity. N-9 and N-11 have a drifting orbit

2004 - 2008 Ozone Monitoring Instrument Full global coverage every day with few missing (OMI) days. OMI produces both ozone and reflectivity values. 1:30 p.m. orbit approximate 13% increase at 50°S latitude near the southern tip of South America. As shown in Fig. 5.12, the percent change in irradiance is dominated by changes in ozone amounts, except near the equator where the change in CT dominates. The apparent leveling off with latitude of ozone change between 35°N and 50°N does not appear as strongly in the irradiance change because of the sec(#) term in the expression for dF12/F2. The increase in cloudiness (decrease in CT) at high southern latitudes moderates the irradiance increases caused by ozone decreases. While these estimated irradiance changes are significant, they are caused by small changes in ozone amount compared to the mean ozone value, which would permit the use of Eqs. (5.1) and (5.2), or the more accurate direct use of Beer's Law (Eq. (5.4)) as shown in Fig. 5.12.

Figure 5.12 The annual average change in annual 305 nm irradiance (solid line) caused by changes (preliminary) in reflectivity and cloud transmission CT (dashed line) and the change in ozone amount O (dot-dash line)

Was this article helpful?

0 0

Post a comment