Showing papers by "Martin Heimann published in 1991"

Three-dimensional simulation of 7Be in a global climate model

Abstract: In the upper troposphere and lower stratosphere, cosmic rays create Beryllium 7 atoms, which subsequently attach to submicron dust particles, so that wet deposition ultimately removes most 7Be from the troposphere. Because this source is weil known and because there is a large climatological data set for 7Be concentration in surface air and deposition on the surface, simulating 7Be provides a good test of the wet scavenging parameterization in a global climate simulation model, such as ECHAM2, which is the European Center for Medium Range Weather Forecasting (ECMWF) model with new physics introduced by the University of Hamburg. A simple parameterization in which the simulated condensation rate determines the scavenging frequency in each grid cell is used in the tracer transport model GLOMAC1, which is embedded in the meteorological model ECHAM2. In this paper we compare observed and model-calculated values of monthly average and annual average surface concentration at a global network of 79 stations. The average absolute value of the error in simulated surface concentration is 1.4 mBq m−3 compared with an average observed concentration of 3.5 mBq m−3. At most stations and in most regions the simulated surface concentration has about the correct magnitude and seasonal cycle, although there is a bias so that the modeled concentration is high at mountain stations in the tropics and low at sea level in polar regions. There are less climatological deposition data than there are climatological concentration data, but the model basically simulates the correct latitudinal variation of the zonally averaged deposition, which has a peak at the polar front (30°–50°N), although the model also has a peak produced by convective precipitation in the intertropical convergence zone, a peak that is not observed. We think that 7Be, used in conjunction with other species such as 210Pb, provides an excellent test of wet scavenging in a global model.

Three-dimensional modeling of the concentration and deposition of 210Pb aerosols

Abstract: Two aerosol-borne species, 210Pb and 7Be, were simulated on-line in a low-resolution global climate model (ECHAM2). This paper mainly considered 210Pb, which basically has its source in the lower troposphere over continents and its main sink in wet deposition. A companion paper (this issue) discussed 7Be, whose source is in the upper troposphere and whose primary sink is also in wet deposition. In order to test a simple wet scavenging parameterization in which the depletion of aerosol in a grid cell depends on the condensation rate in the grid cell, we compared simulated and observed annual average values of 210Pb concentration and deposition: the monthly average surface concentration is generally well simulated; the correlation coefficient is r = 0.79 for annual average concentration and r = 0.63 for annual average deposition, although globally the concentration averages 40% high and the deposition 18% low. Because the scavenging parameters were tuned to give good results for 7Be and reasonable agreement for 210Pb, increasing the rate of wet scavenging would usually improve the simulation of 210Pb but degrade that of 7Be. It is suggested that explicitly considering the cumulus updrafts and scavenging from the updrafts could preferentially scavenge 210Pb relative to 7Be and thus improve the 210Pb simulation without sacrificing the good 7Be agreement. On the other hand, this paper and its companion show that one can often describe the deposition and surface concentration to within about 20% for species with sources in the lower and upper troposphere with a simple “local” (in the vertical) wet scavenging parameterization. Poorer results should be expected, however, if we had tried to match simultaneously the observed concentration and deposition of a species with a true surface source and one with a true stratospheric source.

Parametrization of cloud transport of trace species in global 3-D models

Abstract: It is important to understand the effects of warming of the earth by trace gases and polluting the environment by acid rain. For these and other reasons we need to be able to simulate atmospheric chemistry in the global troposphere. Perhaps the biggest meteorological uncertainty in such modeling is determining vertical transport of trace species by clouds. There has been considerable work on the parameterization of vertical transport of heat and moisture by clouds in meteorological models (see reviews by Anthes et al., 1986 and Frank, 1983); such meteorological parameterizations, however, do not imply a unique tracer transport, and additional assumptions are required. Cloud transfer in global three-dimensional (3-D) tracer models has sometimes been partly resolved but usually is parameterized using a variety of methods, including eddy diffusivities (K), transfer matrices, and convective mass fluxes.

The effect of the global background on a synoptic-scale simulation of tracer concentration

Abstract: We use an Eulerian, three-dimensional tracer model driven by European Center for Medium Range Weather Forecasting winds to study the effect of the global background on synoptic-scale tracer concentration. Inside the global domain we put a synoptic-scale domain that includes North America, the North Atlantic Ocean, as well as parts of the eastern North Pacific Ocean, Africa, and Europe; and inside the synoptic-scale domain we have a regional subdomain over North America. We determine the percentage PI of concentration inside the regional subdomain that came from outside the synoptic-scale domain for species whose only source is at continental surfaces and whose only loss is a spatially and temporally independent first-order decay with an exponential decay time r. In the annual average, PI ≈ 0% at the surface and increases with height: at z = 11.8 km, PI is 3% for τ = 1 day, 50% for τ = 5.5 days, and 60% for τ = 30 days. Month-to-month changes in convection over land and in horizontal transport produce a seasonal cycle in PI with a summer minimum and winter maximum, and the amplitude of this seasonal cycle in PI decreases as τ increases. The most interesting species have τS near 5.5 days, because species with τ ≫ 5.5 days have less of a seasonal cycle in PI and have less variation with height of PI in the free troposphere than for a species with τ ≈ 5.5 days, while species with τ ≪ 5.5 days have such small PIs as to be uninteresting. To help us interpret these Eulerian model results, we use simple Lagrangian models to derive three main conclusions. First, the apparent travel time from other source regions, particularly Asia, to North America decreases a factor of 2 as τ decreases from 30 days to 1 day, which means that more of the 1-day lifetime species arrives than one would expect based on the amount of the longer-lived species that arrives. Second, in the regional subdomain, PI usually increases as τ increases, but sufficiently remote from the sources in the subdomain, PI may decrease as τ increases. Third, although the concentration imported to the middle and upper troposphere generally increases as the vertical mass flux by clouds increases, PI can be insensitive to changes in the cloud mass flux, as long as the cloud mass flux changes similarly both inside and outside the synoptic-scale domain.