Characterization of GOES Aerosol Optical Depth Retrievals during INTEX-APubu Ciren1, Shobha Kondragunta2, Istvan Laszlo2 and Ana Prados3

1QSS Group, Inc., Maryland2 Center for Satellite Applications and Research, NOAA/NESDIS, Camp Springs, Maryland

3UMBC/JCET, Baltimore, MD

The 2004 Intercontinental Chemical Transport Experiment-North America (INTEX-NA) and NOAA’s New England Air Quality Study (NEAQS) field campaigns provided an opportunity to assess the role of satellite data in extending the spatial dimension to study problems such as contribution of long range transport to regional air quality. NOAA/NESDIS provided near real time aerosol optical depth (AOD) product derived from the GOES-12 Imager to aid in the planning of aircraft/ship flight deployment. GOES AODs, available at 30 minute interval over the sunlit portion of the Contiguous United States (CONUS) in 4 km X 4 km spatial resolution, were provided with one hour time lag to the field. The GOES AOD algorithm is a single-channel retrieval based on look-up tables created using a continental aerosol model. However, spatial and temporal variations in aerosol type and size due to varying sources of pollution (e.g., forest fires, urban/industrial, dust) may result in uncertainties in retrieved AOD product. In this study we analyzed GOES AOD measurements with specific focus on uncertainties related to assumptions of one single type of aerosol model across the whole CONUS.

1. INTRODUCTION

2. GOES-M AEROSOL RETRIEVAL

3. GOES AEROSOL OPTICAL DEPTH (AOD) DURING INTEX-A

5. AEROSOL MODELS

5. COMPARIOSNS WITH AIRCRAFT (DC-8) MEASUREMENTS1. ABSTRACT

In this paper, we present results on: (1) the assessment of GOES AOD variations on different temporal (diurnal, daily, and monthly) and spatial (local and regional) scales, (2) validation of GOES AOD with both ground-based (AERONET) and aircraft measurements (DC-8), (3) the effect of varying aerosol model on the retrieved GOES AOD., (4) the applicability of operational GOES AODs in supporting future field campaigns for air quality monitoring, and (5) discuss the potential improvements possible with GOES-R Advanced Baseline Imager (ABI) due to its multiple channels in the visible and infrared.

Current Aerosol Retrieval from GOES

4. COMPARISONS WITH AERONET MEASURESMENTS

INTEX-A study domain and corresponding AERONET sites

a c

bd

In order to be able to retrieve aerosol optical properties, information about aerosol model on both temporal and spatial scale is required. MISR aerosol climatology product is based on classification of the aggregated results from global transport model. It has 13 types of aerosol mixing groups (left figures), based on relative abundance of 4 out of 6 basic components (i.e., sulfate, seasalt, carbonaceous, accumulated dust, coarse dust and black carbon) . The left figures shows the spatial distribution of aerosol models for (a). January; (b). April; (c). July; (d). September.

The 13 MISR aerosol mixing groups (5 major groups, each consists of the same 4 basic components, they are further divided into subgroups according to the percentage of number density of each component) are given as follow:1a. Carbonaceous+ Dusty Maritime 4a. Carbonaceous+ Dusty Continental 1b. Carbonaceous+ Dusty Maritime 4b. Carbonaceous+ Dusty Continental1c. Carbonaceous+ Dusty Maritime 4c. Carbonaceous+ Dusty Continental

2a. Dusty maritime+ Coarse Dust 5a. Carbonaceous+ Black Carbon Continental2b. Dusty maritime+ Coarse Dust 5b. Carbonaceous+ Black Carbon Continental 5c. Carbonaceous+ Black Carbon Continental 3a. Carbonaceous+ Black Carbon Maritime3b. Carbonaceous+ Black Carbon Maritime

The figure on the right shows the simulated GOES-12 ch1 TOA reflectance as a function viewing zenith angle for 13 MISR aerosol types.

1a 1b 1c 2a 2b 3a 3b 4a 4b 4c 5a 5b 5c

6. Effect of varying aerosol model

Background composite image (21 days)

LUT (background AOD 0.04)

Retrieved surface reflectivity

Retrieved GOES AOD (4x4 km), 1/2 hourGOES-12 (East) Visible Image (channel 1)

CLAVR method (GOES-12 IR channels 2 and 4)

LUT

cloud mask

Transport of smoke p lumes from Canada to the US, July 16-20, 2004 as observed by GOES Imager

Vertical profiles of LIDAR aerosol backscatter at UMBC showing that long range transport occurred in the free troposphere between 4 and 5 km

Mean GOES aerosol optical depth during the time period of July and

August, 2004

0.0 0.2 0.4 0.6 0.8 1.0 1.2 1.4 1.60.0

0.2

0.4

0.6

0.8

1.0

1.2

1.4

1.6

R= 0.722 rms= 0.22

Air

craf

t ao

d

GASP AOD

gasp aod (along track) v.s. aircraft aod (dry) gasp aod (along track) v.s. aircraft aod (wet)

Comparison of GOES AOD with AERONET AOD over INTEX-A domain and during the time period from July 1 to August 30, 2004

• Relative error of GOES AOD as a function of scattering angle

• Relative error of GOES AOD as a function of surface reflectance• Scatter plot of GOES AOD against AERONET AOD• Daily mean of GOES AOD (red line) and AERONETAOD (black)

as a function of day of the year• Relative error of GOES AOD as a function of solar zenith angle

0.0 0.2 0.4 0.6 0.8 1.0 1.2 1.4 1.60.0

0.2

0.4

0.6

0.8

1.0

1.2

1.4

1.6

R= 0.711 rms= 0.20

Air

craf

t ao

d

GASP AOD

gasp aod (box) v.s. aircraft aod (dry) gasp aod (box) v.s. aircraft aod (wet)

0.0 0.2 0.4 0.6 0.8 1.0 1.2 1.4 1.60.0

0.2

0.4

0.6

0.8

1.0

1.2

1.4

1.6

R=0.60 rms=0.19R=0.88 rms=0.16

airc

raft

aod

GASP AOD

gasp aod (box) v.s. aircraft aod (wet, aloft) gasp aod (box) v.s. aircraft aod (wet, no aloft)

0.0 0.2 0.4 0.6 0.8 1.0 1.2 1.4 1.60.0

0.2

0.4

0.6

0.8

1.0

1.2

1.4

1.6

R=0.61 rms=0.20R=0.79 rms=0.24

airc

raft

aod

GASP AOD

gasp aod (along track) v.s. aircraft aod (wet, aloft) gasp aod (along track) v.s. aircraft aod (wet, no aloft)

Comparison of aircraft (DC-8) measured AOD with collocated GOES AOD over INTEX-A domain. There are 71 flight tracks during the time period from July 1 and July 24, 2004

a. Scatterplot of the mean GOES AOD against DC-8 measured AODb. Scatterplot of the mean GOES AOD over aircraft flight track against DC-8 measured AODc. Scatterplot of the mean GOES AOD over a box containing aircraft flight track against DC-8

measured AOD, respectively for cases with and without elevated AOD above 3km.

d. Scatterplot of the mean GOES AOD over aircraft flight track against DC-8 measured AOD , respectively for cases with and without elevated AOD above 3km.

e. Vertical profile of Extinction coefficient measured by DC-8. GOES AOD is also shown as a base image

0 20 40 60 80 1000.02

0.04

0.06

0.08

0.10

0.12

0.14

0.16

0.18

0.20

SZA=20, RAZ=40, Srf=0.05, aod=0.64

TO

A r

efle

ctan

ce G

OE

S-1

2 C

H1

Viewing zenith angle

model 1a model 1b model 1c model 2a model 2b model 3a model 3b model 4a model 4b model 4c model 5a model 5b model 5c

a. b.

d.c.

e.

aloft aerosol layer

From the simulated GOES-12 Ch1 TOA reflectance (see left figure), it is shown that differences in TOA reflectance exists between aerosol model 4 and aerosol model 5 for continental aerosol type, however, very small difference is seen among subgroups for each aerosol model. Same feature are shown for maritime aerosol as well. In general, the major difference among aerosol groups is the varying concentration of non-absorbing (i.e., sulfate), less absorbing (i.e., dust) and absorbing (i.e., black carbon) components. The consequent effect on GOES AOD retrieval is clearly shown in above figure. This figure is the same to the figure in the left box, except that biomass burning aerosol model (less absorbing) is applied instead of typical continental aerosol (more absorbing). It is seen that a reduced AOD (about 20%) is produced after applying biomass burning aerosol. Although GOES-12 does not have the ability to distinguish the aerosol, however, by applying prescribed aerosol model climatology, such as the one from MISR, will improve the accuracy of AOD retrieval on a climatologic sense. ABI on board GOES-R – future GOES, will provide calibrated observations over multi-channels with higher refresh rate and spatial resolution, compared to current GOES. These features expand the frequency and coverage of aerosol remote sensing that can support a wide range of weather and environmental applications including the monitoring of air quality.

Mean ratio: smoke/continental aerosol~0.86

Scatterplot of AOD from biomass burning aerosol against AOD from continental aerosol. Note that results shown in this figure are based on the retrieved AOD from the simulated TOA reflectance with original GASP LUT (i.e., continental aerosol) and LUT generated with biomass burning aerosol