Parametric model to estimate clear-sky longwave irradiance at the surface on the basis of vertical distribution of humidity and temperature

The surface downwelling longwave irradiance in clear-sky situations is an important component of the global radiation balance. It can be measured directly using ground-based pyrgeometers or computed using a radiative transfer code given precise information on atmospheric composition (water vapor, ozone, and aerosols) and temperature. Discrepancies between instantaneous observed and simulated values of the clear-sky longwave irradiance are typically at the 3-10 W m^-2 level (root-mean-square error). The discrepancies depend both on pyrgeometer and atmospheric composition uncertainties. Over the past century, many authors have worked on deriving control parameters to simulate the clear-sky longwave irradiance using simple parameterizations. The most common control parameters found in the literature are screen-level temperature, screen-level water vapor density, and column integrated precipitable water. We show that reference parameterizations are able to simulate the clear-sky longwave irradiance with an uncertainty of about 10 W m^-2. Uncertainties are greater during nighttime than daytime periods. We propose a new parameterization that uses the standard input parameters and their diurnal variations to account for important effects of their vertical distribution on the simulation of clear-sky longwave irradiance. The new parameterization allows us to reduce uncertainties in clear-sky surface downwelling longwave irradiance simulations to better than 5 W m^-2 for both daytime and nighttime situations.