“Aerosol” related “613.2’s Image of the Week” from 2007

1. “Aerosol” related “613.2’s Image of the Week” from 2007

2. Interannual Variability of Smoke in Three Biomass Burning RegionsDec 30, 2007 (L.Remer) South America, southern Africa and the Indonesian archipelago represent the three major areas of tropical biomass burning and smoke production. Using MODIS-derived monthly mean Level 3 data, the figure shows a 8 year time series of the aerosol optical depth (AOD) for each of the three regions (right). All three regions show maximum AOD during the July - October season, with southern Africa concentrated in the early part of that period and Indonesia towards the end of the period. The panels on the left compare monthly mean AOD in September and October, 2006 and 2007 for all three tropical biomass burning areas. As reported in Koren et al. (2007) and shown in the Image of the Week September 23, 2007, mean AOD levels in South America increase from 2000 to 2005, then drop significantly in 2006. The reversal of this trend was attributed to a combination of changes in cultural practices and the onset of a early rainy season. However, we see that in 2007, smoke in South America was again at high levels, higher than any other year in the MODIS record. In contrast, 2006 was a significant burning year in Indonesia, while 2007 had relatively little smoke. Interannual variability in biomass burning in Indonesia is linked directly to ENSO cycles, with 2002, 2004 and 2006 all identified as El Nino years. Koren, I., L. A. Remer, and K. Longo, 2007: Reversal of trend of biomass burning in the Amazon. Geophys. Res. Lett.,34, L20404, doi:10.1029/2007GL031530. (Submitted by Lorraine Remer)

3. Remote sensing of cloud sides of deep convective cloudsDec 7, 2007 (T. Zinner) CLAIM 3D -- the three-dimensional cloud and aerosol interaction mission proposal -- is a satellite sensor combination proposed by scientists from NASA’s Climate and Radiation Branch and the University of Maryland Baltimore County. It combines measurements of aerosol characteristics in the vicinity of clouds and profiles of cloud microphysical characteristics. Such a set of collocated measurements will allow new insights into the complex field of cloud-aerosol interactions affecting directly the development of clouds and precipitation, especially in convection. A core instrument is the Cloud Scanner (see October 21, 2007 Image of the Week) which measures radiance reflected or emitted by cloud sides at several wavelengths. A profile of cloud phase and particle size on a high spatial resolution of a few hundred meters will be retrieved from these measurements. For this sensor an experimental retrieval was developed and successfully tested. The retrieval accounts statistically for the complexity of cloud structures and 3D radiative transfer at high resolution. These figures present the test of the proposed retrievals using a completely synthetic test bed. Cloud fields from the Goddard Cumulus Ensemble model (top image) were used as input to a 3D radiative transfer model which simulates the cloud scanner observations (bottom left image). The Bayesian retrieval of particle size distribution was then applied to the simulated data. The retrieved vertical profiles of cloud particle size and cloud phase were compared with the “true” data from the cloud model. The bottom left image shows the simulated observation for the cloud structure above looking on the cloud scene at 60 degrees from the “south” of the scene. The plot on the right shows the true and retrieved cloud properties along the red line in the left image. Mostly cloud sides are observed showing small (up to 12 µm) water droplets and much larger (> 30 µm) ice particles. The results of comparison are very encouraging showing that the method is clearly capable of retrieving the cloud properties along the profile with a high accuracy. The results of this research have been recently submitted to Atmos. Chem. and Phys. Disc. Click here to view the manuscript. There are also two other papers by GSFC Climate and Radiation Branch scientists (to view them, click here and here) that discuss cloud side remote sensing and the proposed CLAIM 3D mission, respectively. (submitted by Tobias Zinner (former Deutsche Forschungsgemeinschaft (DFG) Fellow at GSFC's Climate & Radiation Branch, now at DLR Institut für Physik der Atmosphäre ), Alexander Marshak, and Vanderlei Martins)

4. Changes in summertime U.S. rain with the day of the week and hour of the dayDec 2, 2007 (T. Bell) The image at the top represents the average rain rate during the summers of 1998-2005 for each day of the week and each hour of the day, The averages are over the non-coastal southeast-U.S. as depicted in the map at the bottom. Rain rates are represented by colors; the key to the colors is shown by the color bar in the center. As one moves from the bottom to the top of the color plot, mean rain rates for Saturday, then Sunday, and so on, are shown for each day of the week, then repeated for a second week. Moving from left to right, mean rain rates for midnight through noon (1200 local time) and back to midnight are shown, then repeated for a second 24 hours. A paper now in press by Bell, Rosenfeld, Kim, Yoo, Lee, and Hahnenberger (2008), from which this image is taken, provides evidence that mean summertime rainfall in the SE U.S. is larger in the middle of the week than on weekends. This is believed to be occurring because of the weekly variations in particulate pollution (aerosols). It is well known that in many parts of the U.S. the amount of small particulate matter suspended in the air is larger Monday-Friday than on weekends, probably because of weekly changes in diesel vehicular traffic or power-plant operation. It is also now well established that particulates can affect the way storms develop, and, during the summertime, at least, the effect is to energize storms in this region. The plot shows that, for all days of the week, more rain falls in the afternoons (1200-2400) than during early morning hours (0000-1200), which is typical of summertime rain over land. The plot also shows that it is the afternoon storms that are strengthened midweek, while there seems to be a compensating midweek weakening of morning rain. Why is this important? The effects of aerosols on rainfall and storm development are still being worked out by scientists. They have found that in some environments aerosols tend to suppress rainfall, but in the hot moist environment typical of the SE U.S. in the summer the opposite seems to be the case. Weather forecast models don't yet include the effects of changes in aerosols by human activity, and so they will miss forecasting the changes observed by the satellite. Observations like these will help improve the accuracy of weather and climate models and their forecasts. It is in any case remarkable that, based on the satellite observations, humans are modulating the weather each week over large areas. The rain data used in this plot were provided by the Tropical Rainfall Measuring Mission (TRMM) satellite, which was launched in 1997 and is now celebrating its 10th anniversary of successful operation. TRMM is a joint venture of NASA and the Japanese space agency JAXA. (Submitted by T. L. Bell.)

5. Your eyes can trick you: Where is the cloud base?Oct 21, 2007 (J.V. Martins) The top image is a color composite of blue (410nm), green (550nm) and red (670nm) images on the background of a thick smoke layer. The cloud base is not visible in the smoke background and the smoke itself is confused with the blue sky and the horizon. The two pictures on the bottom contain (a) the visible (550nm) and (b) the near infrared (2100nm) pictures of a piece of the same cloud (red box in the top image) showing the different details covered by both wavelengths. The visible picture (a) shows a cloud structure like we see with our eyes (as the color composite). The 2100nm picture (b) shows two very distinct characteristics as compared to the visible one: 1- The 2100nm picture (b) can see through the smoke and clearly shows a much lower cloud base, which is completely ignored by a naked eye observer or by the analysis of the visible picture. This happens because smoke particles are on average much smaller than 2100nm, producing a weak interaction between light at this wavelength and the smoke. 2- Picture (b) also shows enhanced details in the cloud structure that cannot be observed in the visible one. The structure in 2100nm makes the cloud appear like a volcano or a smoke plume. This characteristic comes from the fact that light at 2100nm is “strongly” absorbed by cloud droplets and on average undergoes much fewer scattering events inside the cloud than light in visible wavelengths. In contrast, visible wavelengths are subjected to “negligible” absorption by the cloud droplets and produce a much smoother picture. In reality, the 2100nm picture shows more realistic details of the cloud dynamic and microphysical structure, which gets lost in the smoothing effect produced by visible light. These pictures were taken in July 2007, during a field campaign in Mount Gibbes/Mount Mitchell, North Carolina, in collaboration with the North Carolina State University. These were the first tests of the NASA Goddard/UMBC cloud side imagers. This system was designed for the study of the interaction between aerosol particles and clouds. There is strong evidence in the literature showing how man made pollution and other aerosols can affect clouds and precipitation. The system also allows for quantitative measurements of the size of the cloud droplets, and for the discrimination of the cloud thermodynamic phase (ice, water, or mixed phase). (submitted by J. Vanderlei Martins)

6. During springtime, Asian dust storms break out frequently in northwestern China, where the Taklimakan and Gobi deserts are located. Mineral dust particulates on such arid surfaces are lifted to higher altitudes (3~7km) by strong wind gusts often associated with springtime frontal systems, and can travel across the Pacific to North America. While Asian dust storms can cause plankton to bloom in the North Pacific Ocean by supplying nutrients such as iron or by reducing the acidity of rain, they also pose significant influences on human health and the climate of the Earth. Dust storms, for example, can carry pollutants (e.g., soot and sulfate) and poisonous metals such as mercury and cadmium as well as potentially harmful microorganisms like fungi and bacteria. Over time, fine dust particles may cause lung disease if inhaled. Dust particles, as effective nuclei, are believed to affect the genesis of ice clouds and their light-reflecting properties, which, in turn, can influence the energy budget of the Earth. The top image, showing two different dust storms over China, was acquired on March 31, 2007 from the MODIS sensor aboard the Aqua satellite. Using the Deep Blue aerosol retrieval algorithm, we can infer the property of aerosols – tiny particles suspended in the atmosphere including mineral dust, smoke, and pollutant particulates. Normally, the quantity of aerosols estimated from satellites is expressed as “Aerosol Optical Thickness (AOT)” as shown in the middle image. It is important to know how much solar energy is absorbed by aerosols as it can influence the climate significantly. A parameter called “Single Scattering Albedo (SSA)” is often used to describe how much solar energy will be scattered or absorbed by aerosols: the lower the SSA, the more absorption. The Deep Blue algorithm is able to estimate SSA for mineral dust aerosols, as shown in the bottom image. Mineral dust from different regions often has different light-absorbing properties. The areas marked in the yellow circle show dust outbreak over the Taklimakan desert while the orange circle delineates dust transported from the Gobi desert. The colors of the two dust plumes are quite different (top): paler for dust from Taklimakan desert, redder from Gobi desert. The quantity (i.e., AOT) of the two dust plumes are more or less similar (middle). According to the Deep Blue retrieval (bottom figure), however, the dust from Gobi desert (orange circle) is more absorbing (SSA 0.88~0.9) at blue wavelengths compared to the dust from the Taklimakan desert (SSA 0.92~0.96). NASA’s continuous efforts to develop, launch and maintain satellites, have given scientists the capability to track and monitor the properties of aerosols from various source regions, including Asian dust storms. The information shown here can be used to help further study how these storms redistribute solar energy, how they affect the properties of clouds, and eventually the climate of the Earth. (Submitted by Myeong-Jae Jeong and N. Christina Hsu) Asian Dust StormsOct 7, 2007 (M.J. Jeong, C. Hsu)

7. The image illustrates the drastic reversal of an increasing trend in biomass burning in just one year, due to a combination of human effort for change and meteorological factors. The upper left panel shows the slopes of linear fits through six years of seasonal mean aerosol optical depth (AOD) as observed by the Terra-MODIS satellite sensor. The regressions were calculated independently for each 1 degree square. The biomass burning season is defined as August - November. The time series spans 2000-2005. We see that smoke increased over the entire Amazon Basin during this period. These trends are as high as 0.05 to 0.1 AOD per year, which represents an increase in AOD of 0.30 to 0.60 over the six year period. Then suddenly in 2006 there was much less smoke. The lower left panel shows the difference in the seasonal mean Terra MODIS AOD between 2006 and 2005. Blues indicate that 2006 had less smoke. The panel on the right shows the interannual variability in MODIS AOD averaged over the entire northern part of South America and also the total number of fire counts summed over the season as observed by AVHRR. We note the tight correlation between total number of fires and seasonal/regional mean AOD. We also note the tightly increasing trends upwards in both data sets until observations in 2006 reverse the trend. Because the smoke was so terrible in 2005 a concerted effort was made by a coalition of governments, scientists and civil authorities in 2006 to monitor burning and mitigate smoke production. Also in 2006 the rains came earlier. The result was dramatic. Smoke from biomass burning is a serious environmental hazard, but unlike earthquakes and severe weather, effective policy can mitigate the severity of the danger to human health, the well-being of the rain forest and the whole climate system. For more information: Koren, I., L.A. Remer, K. Longo, 2007: Reversal of trend of biomass burning in the Amazon. Geophys. Res. Lett., in press. (submitted by Lorraine Remer) Reversal of trend of biomass burning in the AmazonSept 23, 2007 (L. Remer)

8. Lidar solar background light helps studies of aerosol-cloud interactions August 26, 2007 (J-Y. Chiu, A. Marshak, W. Wiscombe) This image shows that solar background light measurements from pulsed lidars can help us better understand interactions between aerosol and cloud. Lidars are commonly used to retrieve vertical distributions of aerosol and cloud layers. The underlying retrieval principle is that the returned signal is proportional to the amount of light backscattered by atmospheric molecules, aerosols and clouds. Measured photon counts are converted into attenuated backscatter profiles, and during the process a number of noise sources need to be accounted for. Solar background light is one of them. Because of limited power, the lidar pulse does not easily penetrate thick clouds, and thus it is widely believed that lidar cloud retrievals (other than cloud base altitude) are limited to optically thin clouds. We have demonstrated that lidars can retrieve optical depths of thick clouds using solar background light as a signal, rather than only noise to be subtracted, as is usually done. The upper panel is a time series of micropulse lidar (MPL) backscatter vertical profiles at NASA/Goddard Space Flight Center on October 29, 2005. Broken clouds were observed, e.g. no cloud or thin clouds at 16:40 and thick clouds at 17:00. The bottom panel shows cloud optical depth retrievals during the same time period, using MPL solar background light (red lines) and AERONET Cimel sunphotometer measurements (blue dots). Validations show that retrieved cloud optical depths agree within 10–15%. More details about this case can be found in MPL-net and this previous image. In short, one can retrieve not only aerosol properties during clear-sky periods via lidar returned signals (active remote sensing), but also the optical depth of thick clouds during cloudy period via solar background lights (passive remote sensing). This indicates that, in general, it may be possible to retrieve both aerosol and cloud properties using a single lidar. Thus, lidar observations have great potential to serve as a unique dataset allowing us to better understand how changes of aerosol in the environment impact cloud properties. The results of this research were recently published in Geosci. Remote Sens. Lett. Click here to view this publication. (submitted by Jui-Yuan Chiu, Alexander Marshak and Warren Wiscombe)

9. What does smoke do to clouds?June 17, 2007 (C. Ichoku) his is a ‘hot’ topic, because smoke is literally hotter than clouds, but also because a number of scientists are looking into this enigmatic issue of smoke (and other aerosol) interaction with clouds. This true-color Aqua-MODIS image acquired on June 14, 2007 at 12:45 UTC, over South-central Africa and the adjoining East Atlantic, portrays a largely cloud-covered scene, with the exception of the lower right quarter, which shows numerous red dots marking the locations of fires. Why is this lower right quarter practically cloud-free, even though it is supposed to have the same land or ocean surface characteristics as the rest of the scene, which is very cloudy? A previous study found that cloud-cover went from about 38% in clean conditions down to 0% in the presence of smoke; thereby demonstrating that smoke inhibits the formation of clouds. This phenomenon may also be linked to the positive impact of smoke on the radiative forcing of climate. A related question is: why is the sky least cloudy directly above the fires than over other areas further downwind where the smoke is transported? A hypothesis that we are investigating is that the heat from large fires probably plays a significant role, not only in inhibiting cloud formation, but also in causing the evaporation of existing clouds directly above the fire. It is common knowledge that fires generate tremendous heat energy, which is transmitted by conduction, convection, and radiation. Conduction is ensured by the ground, convection by the circulation of the air carrying the smoke, while radiative heat is propagated in space by infrared waves. The heat we feel from a fire, even when we are not close to it, is due to the energy transmitted by radiation, and is also what enables satellite sensors to detect fires from space in the first place. Fortunately, MODIS measures the rates of release of Fire Radiative Energy (FRE), which we plan to use to examine the role of the heat from fires in inhibiting cloud formation or in causing the evaporation of existing clouds. (Submitted by Charles Ichoku. Image provided by the MODIS Rapid Response team.)

10. Day-of-week Dependence of Aerosol Concentrations at Rural and Urban EPA Monitoring SitesMay 13, 2007 (T. Bell) Local concentrations of particulate matter smaller than 10 microns, referred to as "PM10", are measured each day at many Environmental Protection Agency (EPA) sites around the country. The image shows for each site whether PM10 concentration exhibits a regular variation with the day of the week, that is, whether PM10 tends to be higher or lower on some days of the week than others. Most of the EPA sites are classified according to their setting as either "rural", "suburban", or "urban and center city". The upper map shows the measured weekly cycle for sites classified as "rural". The lower map shows the same thing for sites classified either as "suburban" or "urban and center city". The maps show, for each site, the day of the week when maximum average particulate concentrations tend to occur, using the directions of the arrows to indicate the day of the week: see the clock-style guide to the directions at the right of the maps. (The foot of each arrow is placed at the site location.) The maps show that particulate concentrations tend to be maximum between Tuesday and Thursday at the vast majority of sites. The lengths of the arrows show the amplitude of the average weekly fluctuations as a fraction of the weekly mean. The key to the meaning of the lengths of the arrows is shown to the right of the color-bar. A fraction f = 0.1 indicates that the average daily PM10 concentration may rise about 10% above the mean on the day it peaks. The colors indicate the likelihood that the weekly fluctuations at that site are truly correlated with the day of the week. A significance level p = 0.1 indicates that there is a 10% chance that a weekly cycle as big as what is seen could be just an accident owing to lack of sufficient data. Thus, blue arrows indicate some uncertainty about the regularity of the weekly cycle, while red arrows indicate fairly strong, regular changes with the day of the week. As one would expect, the weekly changes tend to be stronger in the urban areas than in the rural areas. (Summertime data from 1998 to 2005 were used, although not all site records include all years.) More information can be found in the paper by Bell et al. (2007). (submitted by Thomas Bell)

11. Winter Haze and Fog over the Indo-Gangetic PlainsMarch 18, 2007 (R. Gautam, C. Hsu) Each year during winter season, dense haze and fog engulfs the Indo-Gangetic Plains (IGP), home to nearly half of the total Indian population, for more than a month disrupting the day-to-day life of millions of people. The IGP is one of the most fertile regions of the world and is fraught with high levels of anthropogenic aerosols due to rapidly growing urbanization and industrialization. During winter season (December-January), thick haze and fog often prevail over the IGP due to frequent temperature inversions, low surface temperatures, and enhanced moisture content in the boundary layer. The widespread haze and fog, clearly visible in the satellite image (top left panel), extends over an area of ~1500 km in length and ~400 km in width, with severe fog events blanketing the entire IGP including parts of Pakistan and Bangladesh on the western and eastern sides, respectively. The top right panel shows mean fog occurrences over the IGP, during winter 2000-2006, mapped by meteorological surface observations overlaid with topography data (tallest bar indicates 20 fog occurrence days). Haze particulates in the atmosphere contribute significantly to the formation of fog and cause spatial and temporal variability of fog depending on the aerosol loading and their components such as sulfate, soot, etc. The aerosol loading, indicated by the MODIS aerosol optical thickness (AOT) for the winter season, shows higher values over the IGP compared to the rest of the Indian subcontinent (bottom left panel). The eastern part stands out well in terms of higher AOT due to greater population density and lower topography compared to the entire region. The bottom right panel shows composite of mean fog/low-cloud occurrences for the 6-year winter season derived from the Terra MODIS cloud parameters such as cloud top pressure, cloud fraction and cloud effective radius. The winter haze and fog are likely to have strong coupling with the “aerosol indirect effects” coming into play. For more information, see Gautam, R., N. C. Hsu, M. Kafatos, and S.-C. Tsay (2007), Influences of winter haze on fog/low cloud over the Indo-Gangetic plains, J. Geophys. Res., 112, D05207, doi:10.1029/2005JD007036. (submitted by Ritesh Gautam (GMU) and N. Christina Hsu (NASA).

12. 2006 Annual Mean Aerosol Optical Depth at 550 nm from MODISFebruary 18, 2007 (R. Stockli, L. Remer) The figure shows the annual mean aerosol optical depth at 550 nm (AOD), calculated from daily MODIS 10 km retrievals. The algorithm used to derived these AOD is the new Collection 005 code, which represents a major change from previous collections. Note that the standard MODIS algorithm does not retrieve over clouds, or over snow, ice, sun glint or bright desert surfaces. Thus, the aerosol distribution of the figure is biased toward summer retrievals in high latitude and cloud-free high pressure systems, while totally missing bright desert surfaces. The global aerosol system in 2006 was both typical and atypical of long standing patterns. We see that in 2006 western central Africa and the adjoining Atlantic ocean, the Indo-Gangetic basin in India and northeastern China continue to demonstrate strong aerosol loading as in previous years. Indonesia has unusually high annual mean aerosol loading in 2006. The heavy aerosol loading over eastern Russia and Siberia that appears to intensify over the Arctic Ocean was caused by biomass burning of agricultural lands and boreal forest in Russia during May and transported across Scandinavia and into the Arctic. The MODIS picture is slightly misleading. Because of snow and ice cover, MODIS retrievals are possible only in a limited part of the year. An aerosol event that occurs in the short summer season will be exaggerated in the annual mean values. Only moderate amounts of aerosol are seen over western Europe and eastern North America, even though these regions are thought to produce sulfate and carbonaceous particulate pollution. The Amazon Basin also shows only moderate aerosol loading, although expectations of strong biomass burning would suggest higher values of AOD. Most of the aerosol seen over the ocean in the figure is transported from continental sources. However, the ocean also produces sea salt and dimethylsulfide (DMS) particles. Thus, sometimes land areas report lower values of AOD than surrounding oceans. Image was created by Reto Stockli of the Earth Observatory. Text by Lorraine Remer.