Download Review of the IMPROVE Equation for Estimating Ambient Light Extinction Coefficients PDF

TitleReview of the IMPROVE Equation for Estimating Ambient Light Extinction Coefficients
File Size8.1 MB
Total Pages98
Table of Contents
	Sampling Biases
		Nitrate Sampling Issues
		Ammonium Sampling Issues
		Carbonaceous Aerosol Sampling Issues
		Issues Related to Elemental Species and Coarse Mass
	Chemical Form of Species in the IMPROVE Equation
		Particulate Organic Material (POM)
		Soil Composition
		Coarse Mass
		Sea Salt
	Hygroscopic Growth
Document Text Contents
Page 1

Review of the IMPROVE Equation for Estimating Ambient Light Extinction

J. L. Hand and W. C. Malm


The Interagency Monitoring of Protected Visual Environments (IMPROVE) protocols for

reconstructing ambient light extinction from measured aerosol species is the basis for evaluating

compliance under the Regional Haze Rule. We review the assumptions involved in computing

reconstructed light extinction using the IMPROVE protocol. This review includes examining the

biases in the measurements of aerosol composition, the assumed chemical forms of aerosol

species, particle hygroscopicity, and assumed mass scattering efficiencies. We present a

thorough survey of estimates of mass scattering efficiencies from recent peer-reviewed literature.

Furthermore, we use IMPROVE nephelometry and composition data to estimate mass scattering

efficiencies using a variety of methods. The current mass scattering efficiencies applied in the

IMPROVE equation are then interpreted in the context of this survey and results derived from

the IMPROVE data analyses. Finally, a summary of provisional recommendations for

refinements to the IMPROVE equation and a discussion of important uncertainties to consider in

the assumptions is presented. Although tentative recommendations of refinements to the

IMPROVE equation are presented, final refinements to the IMPROVE equation await future

discussions of the results presented here.

1. Introductio n

The role of aerosols in visibility degradation has been the subject of research for several

decades, but recently interest has intensified with attempts to quantify the optical properties of

aerosols, especially because of the uncertainties surrounding the role of aerosols in climate

change, and because of the need for compliance under the Regional Haze Rule. In most

DRAFT 5/25/2005


Page 2

instances, visibility reduction is primarily due to scattering and absorption by particles. Particle

scattering and absorption properties can, with a number of limiting assumptions, be calculated

using Mie theory (Mie, 1908; van de Hulst, 1981).

Compliance under the Regional Haze Rule is based on protocols for reconstructing

aerosol mass and light extinction (bext) from speciated mass concentrations. Reconstruction

equations are used to estimate PM2.5 mass concentrations (for particles with aerodynamic

diameters less than 2.5 m) as well as light extinction coefficients. Dry PM2.5 fine mass is

computed using equations 1and 2:

PM2.5 = (NH4)2SO4+ NH4NO3 + POM + LAC + Soil (1)

Soil = 2.2Al + 2.49Si + 1.94Ti + 1.63Ca + 2.42Fe (2)

where sulfate is assumed to be fully neutralized ammonium sulfate ((NH4)2SO4), nitrate is

assumed to be in the form of ammonium nitrate (NH4NO3), and organic carbon is included as

particulate organic material (POM), computed by multiplying organic carbon (OC)

concentrations by a molecular weight per carbon weight ratio (POM = Roc·OC). Light-absorbing

carbon is referred to as LAC. We use the term LAC because it is more representative of the

optical properties of light-absorbing carbon rather than elemental (EC) or black carbon (BC),

although these terms are often used interchangeably in the literature. Fine soil concentrations

include the contributions from assumed forms of elemental species (equation 2) (Malm et al.,

1994a). Mass concentrations are given in units of g m-3.

The light extinction coefficient (bext) includes the contributions from light scattering by

particles (bsp) and gases (bsg), and light absorption by particles (bap) and gases (bag):

bext = bsp + bap + bsg+bag (3)

DRAFT 5/25/2005


Page 49

with similar masses. The mass scattering efficiencies of the dry non-hygroscopic and

hygroscopic species are each 3 m2 g-1, respectively, while that of the wet hygroscopic species is,

because of associated water, 8 m2 g-1. When internally mixed, the wet specific mass scattering

efficiency will be between 3 and 8 m2 g-1, depending on the hygroscopic properties of the mixed

aerosol. Several authors discuss aerosol growth characteristics as a function of RH and have

successfully predicted the scattering characteristics of the mixed particles (Anderson et al., 1994;

Sloane, 1983, 1986; Malm and Pitchford, 1997; Malm et al., 2000, 2003, 2005b). However, the

apportionment of scattering to a particular species is still problematic. Typically, one growth

curve is developed for the mixed species, and scattering due to water is apportioned

proportionally among all aerosol components, even though only one of the species may be

hygroscopic. An example would be a mixture of ammonium sulfate and a weakly hygroscopic

organic species.

5.2. Methods for Deriving Mass Scattering Efficiencies

The methods typically used to estimate mass scattering efficiencies depend on the type of

measurements and data available. In this section we will discuss the most common methods

used. We use the term “mass scattering efficiency” to refer to estimates that correspond to a

single aerosol component such as dry ammonium sulfate, and although one could argue that

species such as POM and soil correspond to a combination of chemical components, we will use

this term to correspond to those efficiencies as well. We use the term “specific mass scattering

efficiency” to refer to aerosols composed of more than one species, for example, a “fine mode

specific mass scattering efficiency”. We will also use this term to refer to aerosols that include

water mass. For example, if an efficiency for an inorganic salt particle with associated water

mass is reported, we refer to it as a “specific mass scattering efficiency”.

DRAFT 5/25/2005


Page 50

The simplest method for computing efficiencies is with measured mass concentrations

from filter samples and measured light scattering coefficients from nephelometry. For an aerosol

population, a specific mass scattering efficiency αsp_spec can be defined as the ratio of a light

scattering coefficient corresponding to the mass concentration (M) of that population:


specsp =_α (15)

The average specific mass scattering efficiency can be estimated by dividing the average

scattering coefficient by the average mass concentration for a given set of measurements, or a

linear regression can be performed on the data and the slope of the linear fit can be interpreted as

the specific mass scattering efficiency. Specific mass efficiencies derived in this manner

represent an average aerosol that could be changing due to variations in relative humidity, size

distribution, and composition during the sampling period. Mass concentrations used in this

method could be either gravimetric mass or the sum of chemically analyzed masses, or be

computed from integrated volume distributions using an assumed density.

A number of investigators have taken advantage of the form of equation 13 (bext =

αMjMj) to construct a multi-linear regression (MLR) model with bext as the independent

variable, and the measured aerosol mass concentrations for each species j (Mj) as the dependent

variables. The regression coefficients are then interpreted as mass extinction (or scattering or

absorption) efficiencies. The assumptions required in this formulation are that all the components

contributing to extinction are included, that equation 13 is a reasonable approximation to the

relationship between extinction and the various aerosol species, the number of samples is large

enough to give stable results, and the concentrations of the species are uncorrelated.

However, as stated by White (1991), “Model simulations show the procedure to yield

usefully accurate results under favorable conditions. An important liability is that the availability

DRAFT 5/25/2005


Page 97

Because there is no recognizable spatial pattern in the change in mass scattering

efficiency as a function of mass, it is recommended that constant values of 2.5 and 3.0 m g2 -1 for

inorganic salts and organic mass be assumed; these estimates are consistent with the literature

review and analysis of IMPROVE data. It is further recommended that the other mass scattering

efficiencies in the current IMPROVE equation remain the same and that a dry mass scattering

efficiency of sea salt of 1.0 m2 g-1 be assumed. The following equation reflects these


bext = (2.5)f(RH)[(NH4)2SO4] + (2.5)f(RH)[NH4NO3]+ (3.0)(1.8)[OC]+ (1.0)[Soil]+ (0.6)[CM] + 10[LAC] + (1.0)f(RH)(1.8*Cl)


Discussion of Biases between Reconstructed and Measured Scattering

In this section we present a brief summary of the skill with which each method

reconstructs total scattering as measured by the Optec ambient open air nephelometer. Table 20

presents the average bias (bsp_meas-bsp_recon) for each method (1, 2, and 3) and for the current and

recommended IMPROVE equations. The method 1 analysis yields results which, on average,

have the smallest biases, while method 2 produces results that are only in slightly better

agreement with bsp_meas than those derived from the current IMPROVE equation. Method 3,

which incorporates the relationship between mass and mass scattering efficiency, produces the

smallest average biases. Both the current and recommended IMPROVE equations have larger

average biases than method 3, and the recommended IMPROVE equation has a slightly greater

average bias than the current IMPROVE equation. Because the mass scattering efficiencies for

both the current and recommended IMPROVE equation are not adjusted as a function of mass

loadings, they both will be biased low under higher extinction days and high under lower

extinction days.

DRAFT 5/25/2005


Similer Documents