Collocation mismatch uncertainties in satellite aerosol retrieval validation

Virtanen, Timo; Kolmonen, Pekka; Sogacheva, Larisa; Rodríguez, Edith; Saponaro, Giulia; Leeuw, Gerrit de

doi:10.5194/amt-11-925-2018

Cited by 43 publications

(33 citation statements)

References 32 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Station data, whether AERONET or MAN, is allocated to whichever grid-box they fall in. Point observations will always suffer from representativeness issues (Sayer et al, 2010;Virtanen et al, 2018;Schutgens et al, 2016a), but the representativity of AERONET sites for 1 o × 1 o grid-boxes is fairly well understood (Schutgens, 2019), see also Section 4.…”

Section: Collocation and Analysis Methodologymentioning

confidence: 99%

An AeroCom/AeroSat study: Intercomparison of Satellite AOD Datasets for Aerosol Model Evaluation

Schutgens¹,

Sayer²,

Heckel³

et al. 2020

Preprint

Self Cite

View full text Add to dashboard Cite

Abstract. To better understand current uncertainties in the important observational constraint to climate models of AOD (Aerosol Optical Depth), we evaluate and intercompare fourteen satellite products, representing 9 different retrieval algorithm families using observations from 5 different sensors on 6 different platforms. The satellite products, super-observations consisting of 1° × 1° daily aggregated retrievals drawn from the years 2006, 2008 and 2010, are evaluated with AERONET (AErosol RObotic NETwork) and MAN (Maritime Aerosol Network) data. Results show that different products exhibit different regionally varying biases (both under- and overestimates) that may reach ±50 %, although a typical bias would be 15–25 % (depending on product). In addition to these biases, the products exhibit random errors that can be 1.6 to 3 times as large. Most products show similar performance, although there are a few exceptions with either larger biases or larger random errors. The intercomparison of satellite products extends this analysis and provides spatial context to it. In particular, we show that aggregated satellite AOD agrees much better than the spatial coverage (often driven by cloud masks) within the 1° × 1° grid cells. Up to 50 % of the difference between satellite AOD is attributed to cloud contamination. The diversity in AOD products shows clear spatial patterns and varies from 10 % (parts of the ocean) to 100 % (central Asia and Australia). More importantly, we show that the diversity may be used as an indication of AOD uncertainty, at least for the better performing products. This provides modellers with a global map of expected AOD uncertainty in satellite products, allows assessment of products away from AERONET sites, can provide guidance for future AERONET locations, and offers suggestions for product improvements. We account for statistical and sampling noise in our analyses. Sampling noise, variations due to the evaluation of different subsets of the data, causes important changes in error metrics. The consequences of this noise term for product evaluation are discussed.

show abstract

Section: Collocation and Analysis Methodologymentioning

confidence: 99%

An AeroCom/AeroSat study: Intercomparison of Satellite AOD Datasets for Aerosol Model Evaluation

Schutgens¹,

Sayer²,

Heckel³

et al. 2020

Preprint

Self Cite

View full text Add to dashboard Cite

show abstract

“…However, the spread among the products is slightly more pronounced in summer ( Fig. A1 for JJA, summer for the Northern Hemisphere), when the absolute AOD often reaches its maximum in certain regions (e.g., in China, Sogacheva et al, 2018).…”

Section: Aod Spatial Distributionmentioning

confidence: 99%

“…Similarly, Li et al (2009) concluded that differences in cloud-masking alone could account for most differences among multiple satellite AOD datasets, including several for which different algorithms were applied to data from the same instrument. Due to these discrepancies, none of the satellite AOD products gives identical values of aerosol properties or is uniformly most accurate (de Leeuw et al, 2015(de Leeuw et al, , 2018Kinne et al, 2006). In other words, there is no single "best" AOD satellite dataset globally.…”

Section: Introductionmentioning

confidence: 99%

Merging regional and global AOD records from 15 available satellite products

Sogacheva

Popp

Sayer

et al. 2019

Preprint

Self Cite

View full text Add to dashboard Cite

Abstract. Satellite instruments provide a vantage point to study aerosol loading consistently over different regions of the world. However, the typical lifetime of a single satellite platform is on the order of 5&#8211;15 years; thus, for climate studies the usage of multiple satellite sensors should be considered. This paper assesses some options for creating merged products from an ensemble of 15 individual aerosol optical depth (AOD) data records produced from a broad range of institutions, sensors, and algorithms. Discrepancies exist between AOD products due to differences in their information content, spatial and temporal sampling, calibration, retrieval algorithm approach, as well as cloud masking and other algorithmic assumptions. Users of satellite-based regional AOD time-series are often confronted with the challenge of choosing the appropriate dataset for the intended application. In this study AOD products from different sensors and algorithms are discussed with respect to temporal and spatial differences. Several approaches are investigated to merge AOD records from different satellites, based on evaluation and inter-comparison results. Global and regional comparison of AOD monthly aggregates with ground-based AOD from the Aerosol Robotic Network (AERONET) indicates that different products agree qualitatively for major aerosol source regions on annual, seasonal and monthly time scales, but have regional offsets. All merged regional AOD time series show highly consistent temporal patterns illustrating the evolution of regional AOD. With few exceptions, all merging approaches lead to similar results, reassuring the usefulness and stability of the merged products. Here, we introduce a monthly AOD merged product for the period 1995&#8211;2017, which provides a long-term perspective on AOD changes over different regions of the world. We show that the quality of the merged product is as good as that of individual products. Optimal agreement of the AOD merged product with the AERONET further demonstrates the advantage of the merging of multiple products.

show abstract

“…In brief, this method uses spatial averaging of satellite data and temporal averaging of AERONET data to account for the fact that AERONET provides point measurements of AOD with a sampling frequency of 5–15 min (dependent on instrument configuration) in cloud‐free conditions, while satellites provide an instantaneous swath measurement with individual retrieval footprints covering several to tens of kilometers. This fundamental difference introduces some uncertainty beyond that of AERONET alone, dependent on the level of heterogeneity in the underlying aerosol field (Virtanen et al, ). Since the publication of prior SOAR validation papers (Sayer, Hsu, et al, ; Sayer, Hsu, Bettenhausen, et al, ; Sayer, Hsu, Lee, et al, ), updates have been made to both the AERONET data version and the specifics of the validation methodology, and as a result these are described in full below.…”

Section: Validation Against Aeronetmentioning

confidence: 99%

“…In particular, for example, as noted earlier the bulk of the dust aerosol model matchups is for sites sampling Saharan outflow so error characteristics for other dust‐laden regions may differ (i.e., these metrics may not capture regional variations within a given type). A second limitation is that it folds in some of the uncertainties related to the matchup method and AERONET data rather than the retrieval itself (e.g., Virtanen et al, ). Practically speaking, it would also be reasonable to assume a minimum possible uncertainty of ∼0.01 on the basis of the radiometric uncertainty in the TOA reflectance measurements.…”

Section: Validation Against Aeronetmentioning

confidence: 99%

Validation of SOAR VIIRS Over‐Water Aerosol Retrievals and Context Within the Global Satellite Aerosol Data Record

et al. 2018

View full text Add to dashboard Cite

This study validates aerosol properties retrieved using a Satellite Ocean Aerosol Retrieval (SOAR) algorithm applied to Visible Infrared Imaging Radiometer Suite (VIIRS) measurements, from Version 1 of the VIIRS Deep Blue data set. SOAR is the over-water complement to the over-land Deep Blue algorithm and has two processing paths: globally, 95% of pixels are processed with the full retrieval algorithm, while the 5% of pixels in shallow or turbid (mostly coastal) waters are processed with a backup algorithm. Aerosol Robotic Network (AERONET) data are used to validate and compare the midvisible (550 nm) aerosol optical depth (AOD), Ångström exponent (AE), and fine mode fraction of AOD at 550 nm (FMF). AOD uncertainty is shown to be approximately ±(0.03 + 10%) for the full and ±(0.03 + 15%) for the backup algorithms, with a small positive median bias around 0.02. When AOD is below about 0.2, the AE and FMF have small negative offsets from AERONET around −0.15 and −0.04, respectively. For higher AOD, AE is less offset and the magnitudes of differences versus AERONET are about ±0.2 and ±0.14, respectively. Aerosol-type classifications provided by SOAR are found to be reasonable, matching optical-based classifications from AERONET over 80% of the time. Spatial and temporal patterns of AOD and AE are also compared with those of other contemporary over-water satellite aerosol data sets; dependent on region, the satellite data sets show varying levels of consistency, with SOAR broadly in-family, and the largest discrepancies in regions with persistent heavy cloud cover.Plain Language Summary Aerosols are small particles in the atmosphere like desert dust, volcanic ash, smoke, industrial haze, and sea spray. Understanding them is important for applications such as hazard avoidance, air quality and human health, and climate studies. Satellite instruments provide an important tool to study aerosol loading over the world. However, individual satellites do not last forever and newer satellites often have improved capabilities compared to older ones. This paper evaluates an extension of an algorithm, originally designed to monitor aerosols from an older satellite instrument, to a new satellite instrument called Visible Infrared Imaging Radiometer Suite. The evaluation is performed by comparing to ground truth data which are part of the National Aeronautics and Space Administration's global Aerosol Robotic Network, as well as to other satellite-based aerosol data sets from different spaceborne instruments.

show abstract

Collocation mismatch uncertainties in satellite aerosol retrieval validation

Cited by 43 publications

References 32 publications

An AeroCom/AeroSat study: Intercomparison of Satellite AOD Datasets for Aerosol Model Evaluation

An AeroCom/AeroSat study: Intercomparison of Satellite AOD Datasets for Aerosol Model Evaluation

Merging regional and global AOD records from 15 available satellite products

Validation of SOAR VIIRS Over‐Water Aerosol Retrievals and Context Within the Global Satellite Aerosol Data Record

Contact Info

Product

Resources

About