In situ measurements of snow albedo at five stations along a north–south transect in the dry-snow facies of the interior of Greenland follow the theoretically expected dependence of snow albedo with solar zenith angle (SZA). Greenland Climate Network (GC-Net) measurements from 1997 through 2007 exhibit the trend of modest surface brightening with increasing SZA on both diurnal and seasonal timescales. SZA explains up to 50% of seasonal albedo variability. The two other environmental factors considered, temperature and cloudiness, play much less significant roles in seasonal albedo variability at the five stations studied. Compared to the 10-year record of these GC-Net measurements, the five-year record of MODIS satellite-retrieved snow albedo shows a systematic negative bias for SZA larger than about 55°. Larger bias of MODIS snow albedo exists at more northerly stations. MODIS albedos successfully capture the snow albedo dependence on SZA and have a relatively good agreement with GC-Net measurements for SZA < 55°. The discrepancy of MODIS albedo with in situ albedo and with theory is determined mainly by two related factors, SZA and retrieval quality. While the spatiotemporal structure, especially zonal features, of the MODIS-retrieved albedo may be correct for large SZA, the accuracy deteriorates for SZA > 55° and often becomes physically unrealistic for SZA > 65°. This unphysical behavior biases parameterizations of surface albedo and restricts the range of usefulness of the MODIS albedo products. Seasonal-to-interannual trends in surface brightness in Greenland, and in polar (i.e., large SZA) regions in general, and model simulations of these trends, should be evaluated in light of these limitations.

An understanding of the surface albedo of high latitudes is crucial for climate change studies. MODIS albedo retrievals flagged as high-quality compare well with in situ Greenland Climate Network (GC-Net) measurements but cover too small an area to fully characterize Greenland's albedo in nonsummer months. In contrast, poor quality MODIS retrievals provide adequate spatiotemporal coverage, but are not recommended for use at large solar zenith angles (SZAs) where they have a systematic low bias. We introduce an empirical adjustment to the poor quality data based on high-quality reference albedos and constrained by GC-Net data and theory, and use the adjusted data to improve estimates and fill in gaps of the year-round, Greenland-wide, albedo and surface energy budget. For observations made with SZAs between 55° and 75°, the mean differences (MODIS minus GC-Net) between our adjusted MODIS albedo and GC-Net measurements are −0.02 and −0.03 at Saddle and Summit, respectively, compared to −0.05 and −0.08 between the unadjusted MODIS albedo and GC-Net measurements. The adjusted MODIS snow albedos are usually between 0.75 and 0.87 over dry snow when SZA is larger than 55°, and they reduce unrealistic seasonal and meridional trends associated with MODIS retrievals at large SZA, defined as SZA > 55° and 70°, respectively, for low- and high-quality retrievals. The impact of the adjusted albedo on the surface energy budget, relative to the unadjusted albedo from all MODIS data, is smallest (−0.7 ± 0.1W/m²) in June, and largest (−6.2 ± 0.9 W/m²) in September for the black-sky albedo (BSA). The mean annual absorbed solar radiation (ASR) reduction by the adjusted MODIS albedo in Greenland from 2003 to 2005 is 3.1 ± 0.2 and 4.3 ± 0.2 W/m² for BSA and white-sky albedo (WSA), respectively, about 8.0 ± 0.5% and 10.8 ± 0.4% of ASR based on the raw BSA and WSA. The ASR reduction by the adjusted blue-sky (actual) albedo is between 2.9 and 4.5 W/m², enough to annually melt 27.1 to 41.7 cm snow water equivalent (SWE), or to sublimate 3.2 to 4.9 cm SWE. The ASR difference between the adjusted MODIS BSA and CERES albedo in March from 2003 to 2005 is only −0.1 ± 0.9 W/m², much less than the difference (4.9 ± 1.4 W/m²) between the unadjusted MODIS BSA and CERES. The albedo adjustments exceed the likely direct anthropogenic radiative forcing experienced by Greenland due to greenhouse gases or aerosols. The proposed adjustment preserves most of the zonal and meridional structure of raw MODIS albedo, and extends its usefulness as a cryospheric climate record in times and regions of Greenland with large SZA.

This study analyzes diurnal and seasonal variations of snow albedo at four Baseline Surface Radiation Network stations in the Arctic and Antarctica from 2003 to 2008 to elucidate similarities and differences in snow albedo diurnal cycles across geographic zones and to assess how diurnal changes in snow albedo affect the surface energy budget. At the seasonal scale, the daily albedo for the perennial snow at stations South Pole and Georg von Neumayer in Antarctica has a similar symmetric variation with solar zenith angle (SZA) around the austral summer; whereas the daily albedo for the seasonal snow at stations Barrow, Alaska, and Ny-Ålesund, Spitsbergen, in the Arctic tends to decrease with SZA decrease from winter to spring before snow starts melting. At the hourly scale, each station shows unique diurnal cycles due to different processes that affect snow albedo such as cloud cover, snow metamorphism, surface hoar formation, SZA, solar azimuth angle, and surface features. Cloud escalates the snow albedo at all four stations by shifting solar radiation to visible wavelengths and diminishes the diurnal variation by diffusing incident solar radiation. The 24 h mean snow albedo is higher on cloudy than clear days by 0.02 at the South Pole (December) and Barrow (May), 0.05 at Neumayer (December), and 0.07 at Ny-Ålesund (April). Snow surface structures, for example, wind-channeled sastrugi, appear to be a controlling factor in the diurnal variation of clear-sky snow albedo at the South Pole and Ny-Ålesund. The surface hoar formation cycles and snow metamorphism are consistent with the asymmetric diurnal variation of snow albedo at Neumayer and Barrow. Near the melting point temperature, melt-freeze cycles exceed cloud and surface structure impacts and dominate the diurnal variation of snow albedo at stations Barrow and Ny-Ålesund. The satellite-measured clear-sky snow albedo usually underestimates the average all-sky snow albedo at these stations. These results illustrate the potential biases in daily and monthly albedo products constructed from sun-synchronous satellite daily instantaneous observations which inevitably undersample the diurnal variation of snow albedo.

Objective

We compared the effect of clopidogrel plus aspirin vs aspirin alone on functional outcome and quality of life in the Clopidogrel in High-risk Patients with Acute Non-disabling Cerebrovascular Events (CHANCE) trial of aspirin-clopidogrel vs aspirin alone after acute minor stroke or TIA.

Methods

Participants were assessed at 90 days for functional outcome using the modified Rankin Scale (mRS) and quality of life using the EuroQol-5 Dimension (EQ-5D). Poor functional outcome was defined as mRS score of 2-6 at 90 days and poor quality of life as EQ-5D index score of 0.5 or less.

Results

Poor functional outcome occurred in 254 patients (9.9%) in the clopidogrel-aspirin group, as compared with 299 (11.6%) in the aspirin group (p = 0.046). Poor quality of life occurred in 142 (5.5%) in the clopidogrel-aspirin group and in 175 (6.8%) in the aspirin group (p = 0.06). Disabling stroke at 90 days occurred in 166 (6.5%) in the clopidogrel-aspirin group and in 219 (8.5%) in the aspirin group (p = 0.01). In stratified analysis by subsequent stroke, there was no difference in 90-day functional outcome and quality of life between the 2 groups.

Conclusions

In patients with minor stroke or TIA, the combination of clopidogrel and aspirin appears to be superior to aspirin alone in improving the 90-day functional outcome, and this is consistent with a reduction in the rate of disabling stroke in the dual antiplatelet arm.

Classification of evidence

This study provides Class II evidence that for patients with acute minor stroke or TIA, clopidogrel plus aspirin compared to aspirin alone improves 90-day functional outcome (absolute reduction of poor outcome 1.70%, 95% confidence interval 0.03%-3.42%).