Simultaneous estimation of downward shortwave and longwave radiation, vapor pressure, and diurnal temperature cycles

We have assessed the performance of the MTCLIM scheme for estimating downward shortwave radiation and surface humidity from the daily temperature range, as well as several widely-used schemes for estimating downward longwave radiation, at 50 Baseline Solar Radiation Network (BSRN) stations globally. Biases in all algorithms were relatively small (less than 5% of typical observed values) in most cases. However, average daily downward shortwave had negative biases averaging -40 W/m² (25% of mean daily shortwave) near the coasts, due to the ocean's moderating influence on air temperature; and smaller negative biases of approximately -10 W/m² elsewhere in the presence of snow, which were reduced by the snow correction of MTCLIM 4.3. The largest biases in the humidity algorithm were at rainy coastal sites (generally positive) and at arid continental interior sites (generally negative). Downward longwave radiation was biased upward by an average of 10 W/m² at coastal sites. Elsewhere, biases in downward longwave radiation varied linearly with daily temperature range with a slope of 1.8 W/m²/°C; longwave was relatively unbiased for daily temperature range of approximately 10 °C. A cubic spline interpolation of air temperature to hourly timescales produced plausible estimates of the amplitude of the diurnal cycle in all variables except for vapor pressure and vapor pressure deficit (for which the amplitude was substantially under- and overestimated, respectively), but errors in the shape of the interpolated diurnal temperature cycle introduced negative (positive) biases in the morning (afternoon) in air temperature, vapor pressure deficit, and longwave radiation.