Abstract. Surface full-sky erythemal dose rate (EDR) from the Ozone Monitoring
Instrument (OMI) at both satellite overpass time and local noon time is
evaluated against ground measurements at 31 sites from the US Department of
Agriculture's (USDA) UV-B Monitoring and Research Program (UVMRP) over the period of 2005–2017. We find that both OMI
overpass and solar noon time EDR are highly correlated with the measured
counterparts (with a linear correlation coefficient of 0.90 and 0.88,
respectively). Although the comparison statistics are improved with a longer
time window (0.5–1.0 h) for pairing surface and OMI measurements, both OMI
overpass and local noon time EDRs have 7 % overestimation that is larger
than 6 % uncertainty in the ground measurements and show different levels
of dependence on solar zenith angle (SZA) and to lesser extent on cloud optical
depth. The ratio of EDR between local noon and OMI overpass time is often
(95 % in frequency) larger than 1 with a mean of 1.18 in the OMI product;
in contrast, the same ratio from surface observation is normally distributed with 22 %
of the times less than 1 and a mean of 1.38. This contrast in
part reflects the deficiency in the OMI surface UV algorithm that assumes
constant atmospheric conditions between overpass and noon time. The
probability density functions (PDFs) for both OMI and ground measurements of
noontime EDR are in statistically significant agreement, showing dual peaks
at ∼20 and ∼200 mW m−2, respectively; the latter is lower
than 220 mW m−2, the value at which the PDF of daily EDR from ground measurements
peaks, and this difference indicates that the largest EDR value for a given
day may not often occur at local noon. Lastly, statistically significant
positive trends of EDR are found in the northeastern US in OMI data, but
opposite trends are found within ground-based data (regardless of sampling
for either noontime or daily averages). While positive trends are
consistently found between OMI and surface data for EDR over the southern
Great Plains (Texas and Oklahoma), their values are within the uncertainty of
ground measurements. Overall, no scientifically sound trends can be found
among OMI data for aerosol total and absorbing optical depth, cloud optical
depth and total ozone to explain coherently the surface UV trends revealed
either by OMI or ground-based estimates; these data also cannot reconcile trend
differences between the two estimates (of EDR from OMI and surface observations). Future geostationary satellites with
better spatiotemporal resolution data should help overcome spatiotemporal
sampling issues inherent in OMI data products and therefore improve the
estimates of surface UV flux and EDR from space.