Mapping the Twilight Zone—What We Are Missing between Clouds and Aerosols

Abstract

Scientific understanding of aerosol-cloud interactions can profit from an analysis of the transition regions between pure aerosol and pure clouds as detected in satellite data. This study identifies and evaluates pixels in this region by analysing the residual areas of aerosol and cloud products from the Moderate Resolution Imaging Radiometer (MODIS) satellite sensor. These pixels are expected to represent the "twilight zone" or transition zone between aerosols and clouds. In the analysis period (February and August, 2007-2011), about 20% of all pixels are discarded by both MODIS aerosol and cloud retrievals ("Lost Pixels"). The reflective properties and spatial distribution of Lost Pixels are predominantly in between pure aerosol and cloud. The high amount of discarded pixels underlines the relevance of analyzing the transition zone as a relevant part of the Earth's radiation budget and the importance of considering them in research on aerosol-cloud interactions.

Satellite Observations of Cloud-Related Variations in Aerosol Properties

“…Taken together, the remote sensing uncertainties and systematic aerosol changes near clouds create a dilemma: Excluding observations data from the transition zone in order to avoid remote sensing uncertainties [21] can create a bias toward low aerosol optical depths and weak radiative effects [10,23], but including data from the transition zone despite the remote sensing uncertainties can create a bias toward high aerosol optical depths and strong radiative effects. To resolve this dilemma, we need to improve our ability to measure aerosols near clouds.…”

“…First, this would substantially reduce the available data volume. For example, [21] found that in images taken by the Moderate Resolution Imaging Spectroradiometer (MODIS), 20% of all pixels are deemed too cloudy for aerosol retrievals and not cloudy enough for cloud retrievals. As a result, roughly 20% of pixels are not considered in MODIS-based global estimates of either "clear sky" or "cloudy sky" radiative effects.…”

Mapping and Understanding Patterns of Air Quality Using Satellite Data and Machine Learning

“…Pixels were only incorporated when clear conditions are reported, that is, no contamination of the data due to clouds is to be expected (Lyapustin et al, ). An additional filter was set up to exclude AOD close to clouds to avoid increased AOD near cloud fringes due to aerosol swelling effects (Schwarz et al, ; Várnai et al, ). Therefore, the distance to the nearest cloud as classified by the MAIAC algorithm (Lyapustin et al, ) was calculated and a threshold of 0.1° was set, which corresponds to 7 km.…”

Daytime aerosol optical depth above low-level clouds is similar to that in adjacent clear skies at the same heights: airborne observation above the southeast Atlantic

“…The longer the length, the smaller the number of cloudy-clear pairs, because longer continuous clouds and clear skies are rarer. Furthermore, we set another length, say 20 s, to exclude immediately before and after the edge in order to reduce ambiguity associated with a gradual transition from cloud droplets to unactivated particles, the socalled twilight zone (Koren et al, 2007;Schwarz et al, 2017;Várnai and Marshak, 2018). We convert the temporal dimensions into horizontal ones using the mean true horizontal aircraft speed, which are 200 m s −1 for the ER2 (Fig.…”