Preprint, Proceedings of the A&WMA/AGU Specialty Conference on
Visual Air Quality, Aerosols, and Global Radiation Balance
September 9-12, 1997, Bartlett, New Hampshire

P. B. Russell¹, P. Hignett², L. L. Stowe³, J. M. Livingston⁴, S. Kinne⁵, and J. Wong<sup>6

1</sup>NASA Ames Research Center, Moffett Field, CA 94035
²United Kingdom Meteorological Office, Meteorological Research Flight, DRA Farnborough, Hampshire, GU146TD, UK
³NOAA/NESDIS, Climate Research And Applications Division, NSC, Washington, DC 20233
⁴SRI International, Menlo Park, CA 94025
⁵Bay Area Environmental Research Institute, San Francisco, CA 94122
⁶Atmospheric Science Program, Department of Physics, Dalhousie University, Halifax, Nova Scotia, Canada, B3H-3J5


Abstract. TARFOX measured a variety of aerosol radiative effects (including direct forcing) while simultaneously measuring the chemical, physical, and optical properties of the responsible aerosol particles. Here we use TARFOX-determined aerosol, gas, and surface properties to compute radiative forcing for a variety of aerosol episodes, with midvisible optical depths ranging from 0.07 to 0.6. We calculate forcing by several techniques with varying degrees of sophistication, in part to test the range of applicability of simplified techniques--which are often the only ones feasible in climate predictions by general circulation models (GCMs). We then compare computed forcing to that determined from aircraft measurements.

The calculations and measurements all yield instantaneous daytime aerosol direct radiative forcing in the range -50 to -190 W m^-2 per unit midvisible optical depth. The magnitudes are about 15 to 100 times the global-average direct forcing expected for the global-average sulfate aerosol optical depth of 0.04. The reasons for the larger forcing in TARFOX include the relatively large optical depths and the focus on cloud-free, daytime conditions over the dark ocean surface. These are the conditions that produce the actual major radiative forcing events that contribute to any global-average climate effect.


1. Introduction

Radiative forcing is defined as the change in the net (downwelling minus upwelling) radiative flux at a given level in the atmosphere [IPCC, 1995]. This net flux is the radiative power density available below that level to drive climatic processes in the earth-atmosphere system. Recent research shows that radiative forcing by aerosol particles is a major source of uncertainty in climate predictions. To reduce those uncertainties, TARFOX was designed to determine direct (cloud-free) radiative forcing by the aerosols in one of the world's major industrial pollution plumes--that flowing from the east coast of the US over the Atlantic Ocean.

TARFOX measured a variety of aerosol radiative effects (including direct forcing) while simultaneously measuring the chemical, physical, and optical properties of the aerosol particles causing those effects. The resulting data sets permit a wide variety of tests of the consistency, or closure, among the measurements and the models that link them. Because climate predictions use the same or similar model components, closure tests help to assess and reduce prediction uncertainties.

In this work we use the TARFOX-determined aerosol, gas, and surface properties to compute radiative forcing for a variety of aerosol episodes, with midvisible optical depths ranging from 0.07 to 0.6. We calculate forcing by several techniques with varying degrees of sophistication, in part to test the range of applicability of simplified techniques--which are often the only ones feasible in climate predictions by general circulation models (GCMs). We then compare computed forcing to that determined from aircraft measurements.

2. Measurements

In this paper we use the following measurements.

2.1 Radiative Fluxes (Irradiances)

The aircraft forcing measurements were made by the UK Meteorological Office Meteorological Research Flight C130. Shortwave upwelling irradiance and shortwave downwelling direct and diffuse irradiances were measured in cloud-free conditions over a range of total column aerosol loadings and over the depth of the aerosol layers. These data cover the 0.3 to 3.0 micron and 0.7 to 3.0 micron wavelength ranges and are used in the
calculation of the visible column optical depth and the direct radiative forcing induced by the aerosol [Hignett and Taylor, 1997].


2.2 Aerosol Properties

The UK C-130 also measured the accumulation mode aerosol size spectra. In addition the University of Washington C-131A made a wide variety of measurements [Hobbs, 1996, 1997], of which we use the following:

• Aerosol composition, showing the relative amounts of aerosol water, carbonaceous material, and sulfate [Novakov et al., 1997a,b; Hegg et al., 1997a,b]
  
  
• Aerosol scattering and absorption, which yielded the dry aerosol single-scatter albedo and bounds on the wet single-scatter albedo [Hegg et al., 1997a,b]
  
  
• Aerosol optical depth spectra [Livingston and Russell, 1997].

We also use analyses of skylight measurements reported by Remer et al. [1997] as an additional means of bounding single-scatter albedo values.

2.3 Absorbing Gases

The radiative calculations require water vapor and ozone vertical profiles or column contents. For these we use a combination of aircraft measurements, satellite measurements, and climatological values.


2.4 Surface Albedo

Surface albedo values are derived from a large set of over-ocean C-130 measurements reported by Glew et al. [1997]. In particular, we use the fitting equation reported there to describe the dependence of ocean surface albedo on solar zenith angle.


3. Calculation Methods

We calculate radiative forcing by three methods, which we call Methods 1a, 1b, and 2.

3.1. Methods 1a and 1b

Figure 1 is a schematic illustration of the sequence of calculations in Methods 1a and 1b. These methods rely on aerosol layer size distributions derived from optical depth spectra. As illustrated in the upper frames of Figure 1, the size distributions are derived using the constrained linear inversion technique of King et al. [1978]. The derivation requires the real and imaginary refractive index at the wavelengths of the optical depth measurements (380, 453, 525, 1020 nm [Livingston and Russell, 1997]).

3.1.1. Method 1a and 1b Refractive Index Model
In Methods 1a and 1b, the refractive index spectra are estimated in the following manner. We use the results of Hegg et al. [1997a, b] that:

• Water was the dominant contributor to TARFOX C-131 optical depths (contributing, on average, 51% of aerosol optical depth for >0.2)
  
  
• Carbonaceous material (absorbing and non-absorbing) was next most important contributor, with sulfates a close third.
  
  
• The mean dry single scatter albedo was 0.90, and the mean upper limit for wet single scatter albedo was 0.94 (with an upper limit range of 0.87 to 0.97).

We also use the result of Remer et al. [1997] that

• The best fit to skylight scattering measurements was obtained by using a wet single-scatter albedo of 0.96-0.97.

In the above, the terminology wet describes aerosol particles in the state of hydration they maintain at ambient conditions; dry describes the particles after they are dehydrated by the sampling techniques of Hegg et al. [1997ab].

In Methods 1a and 1b, the above results are used to develop an approximate refractive index model. The model uses the Palmer and Williams [1975] results for 62% water/38% sulfuric acid, with imaginary indices m_i modified as follows. For all wavelengths where the Palmer and Williams m_i <0.005, we increased m_i to the value that gave midvisible single-scatter albedo (550 nm)0.96. For most size distributions used, this gave m_i = 0.005. The rationale for choosing the 62% water/38% sulfuric acid proportions in the Palmer and Williams results is that these are the Palmer and Williams proportions that most closely approximate the Hegg et al. water/sulfate proportions ([51%/27%]/[51%+27%] = 65%/35% in the absence of carbonaceous materials).

The approximate refractive index model of Methods 1a and 1b does not account for the carbonaceous material reported by Novakov et al. [1997a,b] and Hegg et al. [1997a,b] except by adjusting imaginary refractive indices to obtain the wet single-scatter albedo of 0.96 noted above (which is consistent with the range of wet single-scatter albedos described by Hegg et al. [1997ab]).

3.1.2. Method 1a and 1b Radiative Calculations

Using the refractive index model of Methods 1a and 1b in the King et al [1978] constrained linear inversion yields an estimated size distribution for each aerosol optical depth spectrum, as exemplified by the upper frames of Figure 1. This size distribution is then used with the refractive index spectrum at eight wavelengths across the solar spectrum to calculate the corresponding single-scatter albedo , scattering asymmetry parameter g, and optical depth . An example result is shown in the lower left frame of Figure 1. Having , g, and across the solar spectrum permits calculation of aerosol radiative forcing--i.e. the change in the net (downwelling minus upwelling) radiative flux caused by the aerosol layer.

Above an aerosol layer the dominant term in the aerosol radiative forcing is the increase in upwelling flux [Russell et al., 1997a]. Method 1a uses the simplified result given by Eq. (5') of Russell et al. [1997a] to calculate this flux increase. Method 1b uses the aerosol layer transmission and reflection equations called Model 1 by Coakley and Chylek [1975] in the multiple-reflection equation of Russell et al. [1979]. Example results are shown in the lower right frame of Figure 1.

Below an aerosol layer overlying a low-albedo surface, the dominant term in the aerosol radiative forcing is the decrease in downwelling flux. Method 1a does not include a result for this decrease, but Method 1b does. Specifically, Method 1b calculates the decrease in downwelling flux below the layer by using the aerosol layer transmission and reflection equations called Model 1 by Coakley and Chylek [1975] in a multiple-reflection result derived by Chylek and Wong [1997].
Results of Methods 1a and 1b are described in Section 4.

3.2. Method 2

Method 2 uses accumulation mode aerosol size spectra measured by the UK C-130 together with a refractive index model different from that used in Methods 1a and 1b.

3.2.1. Method 2 Refractive Index Model
Like the Method 1 refractive index model, the Method 2 refractive index model is based on the composition results of Novakov et al. [1997a,b] and Hegg et al. [1997a,b] cited above. However, the Method 2 model uses those results in the program of Lowenthal et al. [1995] to compute explicitly the real and imaginary refractive index spectra. That program assumes that the aerosol particles are an internal mixture of the measured constituents and computes refractive indices from the indices of the constituents and their relative amounts.

3.2.2. Method 2 Radiative Calculations

Method 2 uses a new flexible radiative transfer model [Edwards and Slingo, 1996: Taylor et al.,1996]. For this application the solar spectrum is divided into 220 bands and a delta-Eddington routine used for the calculation of radiative fluxes. The direct aerosol radiative forcing is then defined to be the difference between the case with aerosol particles present and that without. The vertical resolution is approximately 10mb with full account taken of multiple reflection between the aerosol layer and overlying atmosphere, as well as between the aerosol and the surface.


4. Results

We have calculated aerosol radiative forcing for several days on which TARFOX measurements of aerosol properties and effects were made. Table 1 lists the dates for which Method 1a and 1b calculations have been made. Note that these cases include a wide range of midvisible optical depths [() = 0.06 to 0.6] and wavelength dependences.

Figure 2 shows results of Method 1b calculations for the change in upward flux (termed upward forcing) above the aerosol layer and the change in downward flux (termed downward forcing) below the layer. These results are expressed as forcings per unit optical depth and shown as a function of the cosine of solar zenith angle . The day-to-day difference in results is caused primarily by differences in size distribution and wavelength dependence of optical depth (the former being derived from the latter in Method 1). The different size distributions produce somewhat different single-scatter albedo spectra (), though for all distributions (500 nm) is between 0.95 and 0.97. Some of the day-to-day difference in forcings per unit optical depth is also caused by optical depth differences, because the forcing is not exactly a linear function of optical depth (see, e.g., Russell et al. [1997a] and below).

The upward and downward forcings in Figure 2 appear as mirror reflections of one another, except that the magnitude of the negative downward forcings exceeds that of the positive upward forcing by about 40 W m^-2. This difference is caused by absorption in the aerosol layer resulting from the nonunity single-scatter albedo values.

Figure 3.

(a) Cosine of solar zenith angle, hemispheric upscattering coefficient for wavelength 0.7 mm, and surface albedo, all for 25 July, 39.0 N, 75.1 W. (b) Change in upwelling flux at top of aerosol layer derived for several TARFOX days using Methods 1a and 1b.

Figure 4

Comparison between aerosol downward radiative forcing determined from C-130 irradiance measurements (data points) and calculated by Method 2 for several aerosol compositions.

Figure 5.

Comparison between aerosol downward radiative forcing determined from C-130 irradiance measurements (data points) and calculated by Method 1b for size distributions retrieved from sunphotometer optical depth spectra for two days.