Giovanni visualizations generated by the user community
This page provides Giovanni imagery generated by the user community for examination and discussion
The Giovanni images shown here are generated by members of the worldwide Giovanni user community or by the GES DISC staff. The images are posted here for examination and discussion. Members of the Giovanni Facebook group can discuss these images; anyone can send images and figures to the page editor for posting here.
Please provide a caption and brief description of the related research, or the research questions that may be related to the image, to provide context and engender discussion. You may include your name or indicate that you wish to submit anonymously.
Send imagery and related text to firstname.lastname@example.org
Earth Observatory posted an article about this tragic and massive flooding event:
So I decided to see what I could do in Giovanni with a customized palette. I used the Ocean palette. I thought this provided a good perspective on the size of the system, with the red borders showing how large the system was, and the light blue giving a good indication of where the heaviest rain was. I looked at just May 1-2.
The article accompanying the "Image of the Day" of North American snow cover said the following:
"According to NOAA, “Across North America, snow cover for April 2010 was 2.2 million square kilometers below average—the lowest April snow cover extent since satellite records began in 1967 and the largest negative anomaly to occur in the 521 months that satellite measurements are available.” Unusual warmth descended on North America in April, leading to both low snowfall amounts and rapid melt of existing snow."
As of May, Giovanni doesn't have any snow cover products available yet; they should be in MERRA and GLDAS eventually. So the image below was generated from NEESPI Daily . It is an image of AIRS daily surface temperatures (ascending), averaged for April 2010, over North America. (Units are degrees K; 260 K is -13° C or 8.6° F; 288 K is 14.85° C or 58.7° F) The area of warm temperatures in the Canadian Great Plains corresponds almost exactly to an area of zero snow cover in the Earth Observatory Image of the Day.
The NASA Earth Observatory Image of the Day on June 8, 2010 was a visualization of the precipitation and storm track of Tropical Cyclone Phet , which arose in the Arabian Sea and hit the easternmost coast of Oman, causing damage and fatalities, in the first few days of June 2010.
Because we have the same data set in Giovanni that was used by the Earth Observatory – the TRMM Multi-Satellite Precipitation Analysis Near-Real-Time Monitoring product – I decided to see if I could come up with a visualization of Phet's rainfall to compare with (and maybe have a larger WOW! factor than) the Earth Observatory visualization.
My results are below. On top, the Purple-Red-Stripes palette was used. This palette gives a nice contour effect. In the middle, the Ocean palette was used, which also has a contour effect and which brings out areas of lighter rainfall. At the bottom. the BLUE-RED (not the blue-red) palette was used. The upper bound value for rainfall was adjusted downward slightly for all of these images. See my additional comments below the third image.
The intense interest in the Deep Horizon catastrophic oil spill in the Gulf of Mexico has generated interest regarding whether or not the ocean color radiometry data products acquired from the Ocean Biology Processing Group (OBPG) can be used to detect it. The data products from the OBPG in Giovanni are monthly products; thus, they are averaged over the entire month. The daily imagery from the Moderate Resolution Imaging Spectroradiometer (MODIS) on Terra and Aqua has observed the oil spill in the "sun glint" regions of the MODIS scanning swath, where the oil spill affects the reflection of light from the ocean surface in the area where sunlight is directly reflected back to the instrument. In ocean color radiometry data, the sun glint area is masked out, because the direct reflection area does not provide useful data -- ocean color radiometry uses light that has been absorbed and scattered below the surface, not at the surface. Because MODIS looks directly down (nadir observations) from the satellite, its data will have a broad sun glint area.
In contrast, the Sea-viewing Wide Field-of-view Sensor (SeaWiFS) can tilt, which means that by making observations at an angle, it collects light that is not reflected directly back upward to the instrument. Thus, it can observe the area that for MODIS is affected by sun glint. Chlorophyll concentrations due to absorption by chlorophyll in phytoplankton are determined by ratios of light intensity for the different observational bands in the instrument because light on the blue side of the visible spectrum is preferentially absorbed by chlorophyll. On the red side of the spectrum, the light is only reflected by whatever is in the water; this means that the bands on the red side, particularly the band at 555 nm, is good for detecting non-absorbing substances and particles.
We worked with Stan Terzopolos, Coordinator for the Eastern Region of the Pennsylvania PIMS, Partners in Mathematics and Science Program (who accompanied Pennsylvania teachers visiting Goddard Space Flight Center to a "hands-on" training session), to see if there was any detectable signal of the Deep Horizon oil spill in the ocean color radiometry data in Giovanni. Chlorophyll did not show an apparent signal; examination of the new Colored Dissolved Organic Matter (CDOM) product seemed promising, but the data are too diffuse to show an obvious signal from the oil-affected waters. So we also examined the remote sensing reflectance at 555 nm data (Rrs555).
This data shows an interesting distribution in June 2010. Comparing to the MODIS images, the regions with elevated Rrs555 in in June appear to correspond to the general pattern that is seen in MODIS pictures from May and June. We know, however, that oil is a very dark, non-reflective substance. So is this a signature of oil -- and why is it much more evident in June than in May? What could be occurring is that as the oil "ages", it becomes lighter in color, and this lighter-colored oil (media pictures show that it is orange-ish) may reflect sufficient light on the red side of the visible spectrum to be observed in the SeaWiFS data. The June image suggests that the Rrs555 data is detecting some of the Deep Horizon oil in the Gulf of Mexico.
Below are two MODIS images from May 24, 2010 and June 25, 2010, and two SeaWiFS Rrs555 images for May and June 2010.
(above) MODIS image of the Gulf of Mexico, Mississippi Delta region, June 25, 2010, showing the reflection of the oil from the Deep Horizon oil spill in the MODIS sun glint region. Labels show the location of Mobile, Alabama, and the approximate location of the Deep Horizon oil rig.
In an article on TerraDaily, "Black Carbon Implicated in Global Warming" , the file MODIS image accompanying the story looked a lot like a Giovanni plot (though it isn't; it was likely created with IDL). It was fairly easy to generate the same regional plot with the Giovanni MODIS Monthly interface (shown below); the palette maximum was adjusted slightly so that it would look more like the file image in the article. The article summarizes the results of a paper published in Nature Geoscience, " Warming influenced by the ratio of black carbon to sulphate and the black-carbon source ", by Ramana et al. – the second author is the esteemed V. "Ram" Ramanathan of Scripps Institution of Oceanography. In the paper, the authors indicate that the most net warming from the emissions occurs when the black carbon-to-sulphate ratio is high, and that emissions strategies intended to help reduce global warming should seek to both reduce black carbon and the black carbon-to-sulphate ratio. (By the way, this is a MODIS Terra image reproducing the TerraDaily image, and we also have MODIS Daily data from Terra. How confusing!)
At the recent Earth System Science Education Alliance ( ESSEA ) annual conference in Ithaca, NY, a session with education module developers included an examination of the Atmospheric Infrared Sounder (AIRS) carbon dioxide (CO2) data product over a spatial region – which was not something we had seen before. Paul Adams of Ft. Hays State University in Hays, Kansas, led the discussion and examination of the data. This discussion induced the generation of the Giovanni visualizations that are shown below, and provided several "revelations" about the data, the dynamics of CO2 in Earth's atmosphere, and even the Giovanni system itself.
The first visualization created was the AIRS CO2 data product over the United States (and adjacent North American regions) in July 2005, which is the third image in the sequence below. This image (once we figured out how to interpret it), showed some interesting patterns. Over forested areas (such as the Pacific Northwest and northern Rocky Mountains, the central Great Lakes region, the U.S. Northeast, and most of Canada), CO2 concentrations were low, indicating continuing uptake by plants, mostly info forests. In contrast, in the Great Plains and desert Southwest, CO2 concentrations were higher, indicating less uptake by plants – the growing season in the agricultural fields of the Midwest is nearly over in the warm days of July. The data have remarkably good spatial resolution.
Following the meeting, plots of CO2 for each July (2003 - 2010) were created. Initially, it was thought this would be easy for Giovanni, by just changing the "Begin Year" and "End Year" fields to generate a plot for July of each year. But the plot colors kept changing, so the pattern seen in 2005 wasn't as obvious in other years.
The reason for this is the way that Giovanni generated the color plots. The default color palette option is a "dynamic" scale, depending on the range of data values. Because the mininum and maximum range was different each year, the plots didn't look the same. So a standard customized palette was substituted. But when the years changed, the appearance of the plot changed, too. The reason this time was that atmospheric CO2 concentrations are increasing, so a color palette range that worked for July 2003 didn't cover the same range of data values for July 2009.
In order to portray each successive July with approximately the same color palette range, the range of minimum and maximum values had to be adjusted upward (a "moving window"). The range was usually 5-6 ppm. In the plots below, the range is shown above the plot. For July 2003, the range is 373-379 ppm; for July 2010, the range is 389-393 ppm.
One of the capabilities that would be a desirable addition to Giovanni is making an average of a data product for a given time period (usually a month) over several successive years – in essence, a climatology for that time period, if the period of time over which the data are averaged is long enough. This example, however, shows that one of the assumptions of this process would be a flat baseline, i.e., the data would fall in about the same range of values, year after year. In the case of atmospheric CO2, that isn't possible, because the atmospheric concentration is increasing every year. The annual change (a "sloped" baseline) would have to be subtracted out to provide a depiction of the variability of CO2 concentrations over North America (or another region) over several years in July, or a different month, or a season.
In the images below, one of the striking differences was in July 2009. The summer of 2009 in the eastern part of the United States was cooler and wetter than normal; this appears to have enabled more growth in the western Great Lakes states (Indiana, Ohio. Pennsylvania) and particularly in eastern Canada. Homeowners know that a summer alternating rain with cool sunny days means a lot more lawn-mowing, because the grass keeps growing. It appears that the forests and crops kept growing in 2009, too.
As is often the case, the wonderful late summer and autumn weather in much of the United States (especially on the East Coast) has an underlying downside – low rainfall and drought conditions. That's happening again this September. One major way that these conditions are somewhat alleviated is when a hurricane or tropical storm comes inland. Usually hurricanes and tropical storms are not welcome on the coast, but their rainfall amounts (provided they aren't too excessive) can be beneficial.
Tropical Storm Hermine illustrates this point. The rain from Hermine covered a broad area from Texas all the way to Illinois, as shown in the TRMM rain track plot below. (Note: this plot was generated with our 2nd-generation Giovanni, as our hardware maintenance activities have affected data update on other Giovanni machines.)
The Tropical Rainfall Measuring Mission (TRMM) acquired rain rate data for the strong line of thunderstorms that passed through the upper Midwest of the United States on November 22nd. The storms caused tornadoes that were sighted in Illinois and Wisconsin. The tornado may have overturned a school bus near Caledonia, Illinois; six children were taken to the hospital. Observers took pictures and video of the tornado on the ground near Rock Cut State Park outside of Rockford, Illinois. In Wisconsin, heavy damage near Union Grove (west of Racine, near the source of the Des Plaines River) was attributed to a possible tornado.
"Surprise" Tropical Cyclone Carlos dumped record rains on the northwest coast of Australia and the city of Darwin over the past week . At the time that this note is being written, Carlos is still located on the northwest coast of Australia, near Onslow and Exmouth. It is expected to move out to sea and will no longer be a hazard to the Australian citizens.
The Earth Observatory frequently provides users with inspiration for the use of Giovanni, and an early spring dust storm in Iraq in the first days of March 2011 was one of those events. The Image of the Day article entitled "Dust Travels from Iraq to Iran" showed true color images of the dust storm on March 3rd and March 4th. To look at this event, aerosol optical depth (AOD) images for the period March 2-5 were generated, to provide more data in the image than would be seen in a single day.
The results are interesting. North of the large and dark Mileh Tharthar wetland, the storm is clearly seen in the true color images on March 3. But the MODIS AOD images do not show significantly elevated AOD in that location. Instead, higher AOD values are seen in southeastern Iraq. (Mileh Tharthar is the northern of the two large central lakes in Iraq that can be seen in the Giovanni images.)
As these are images averaging all the data for this region over March 3-5, it is likely that when the storm was near ground level and the dust was thick on March 3, it was interpreted as land. The dust likely mixed higher into the atmosphere on subsequent days as it was transported to the southeast, resulting in the elevated AOD values seen in eastern and southeastern Iraq, southern Iran, and over the Persian Gulf.
MODIS-Terra Aerosol Optical Depth, March 3-5, 2011
NASA' s Earth Observatory has true-color Moderate Resolution Imaging Spectroradiometer (MODIS) images of dust storms that occurred on March 26-27 and April 13; the April 13 storm is the one described in the linked article.
Videos of the event ( such as this one ) show daytime transformed to nighttime as the thick cloud of dust descended over the countryside.
A persistent, slow-moving low pressure system brought a lot of rain to the Mid-Atlantic region (relatively speaking, compared to the Midwest's rain and flood disaster) in mid-May 2011. The Potomac River was also in flood stage on May 19 and 20, with flooding on the upper tidal Potomac River, and the Georgetown flood walls were raised. (Earlier in the year when rains caused the Potomac to flood, the walls weren't raised, causing damage to some riverside buildings and restaurants).
The first TRMM accumulated rainfall map shows that most of the rain fell over West Virginia and Virginia. Several storm bands moving west-northwest dropped a lot of rain on central Virginia. There was so much rain in central Virginia that it looks like most of the state of Maryland didn't get any!
TRMM 3B42RT accumulated rainfall map, May 16-20, 2011, over the Mid-Atlantic region
Because here Greenbelt (where Goddard Space Flight Center is located) and central Maryland there was definitely some rain, Giovanni was used to zoom in on the Chesapeake Bay region. It still looks like there wasn't much rainfall in the upper Chesapeake Bay region around Baltimore and Havre de Grace. (Baltimore is located at approximately 39.3 degrees N, 76.6 degrees W. The reason for this appearance is the higher precipitation values at the edges of the image. Giovanni dynamically adjusts the color palette to cover the full range of data values.
On March 22, 2010, a dust storm that had originated in the Gobi Desert and north-central China swept through Hong Kong and out to sea, causing widespread air pollution and haze in the city of Hong Kong, the island of Taiwan, and parts of Japan. A Giovanni user who works in Hong Kong said that this episode is one of two episodes under study, and "it is the worst episode recorded in Hong Kong so far". The Vertical Feature Mask visualization using CALIPSO lidar data was used to show the vertical distribution of dust aerosols in the atmosphere for this event. Using the latitude/longitude scale under the image, which represents a vertical "slice" of the atmosphere along the CALIPSO ground track, the location of Hong Kong can be found at 22.3° latitude, 114 longitude. Fortuitously, CALIPSO went directly over Hong Kong on this date and time. The yellow color of features identified as aerosols is mixed with the blue color of features identified as clouds.
News reports and images show the it was very cloudy and hazy in Hong Kong on March 22:
Vertical Feature Mask plot for 22 March 2010, showing dust in Hong Kong and central China (generated by Giovanni 3 using data from CloudSat). Click on the image to see it larger.
A Wunderblog post by Dr. Jeff Masters on June 14, 2011 described the National Oceanic and Atmospheric Administration's (NOAA) report that the meteorological spring of 2011 was the wettest on record for the continental United States. (Meteorological spring is the months of March, April, and May, while astronomical spring is the period between the spring equinox and the summer solstice.) The NOAA report gave the statewide rankings in terms of how much precipitation each state received, but did not provide actual numbers. Using the Tropical Rainfall Measuring Mission 3B42RT Near-Real-Time monitoring product, an estimate of the actual amount of precipitation can be made using data acquired by satellite-borne instrumentation:
Just for fun, we put an open pitcher of water on the back deck of our house in Elkridge, Maryland to see how much rainfall we received. We measured the accumulation at about 5.25 inches (133 millimeters). The image below shows the TMPA-RT data product for the states of Maryland and Delaware, and parts of North Carolina, Virginia, West Virginia, Pennsylvania, New Jersey, and New York. The location of Elkridge, Maryland (just south of Baltimore) marked. The color scale indicates that the Elkridge region received between 120-140 millimeters of rain – right in the correct range for the amateur measurement we made. To "correlate" this measurement, the official reported accumulated rainfall from Irene at Baltimore Washington International Thurgood Marshall Airport was 4.69 inches (119 millimeters). So give-or-take a few millimeters, the TMPA-RT product appears a pretty accurate estimate of Irene's rainfall in Maryland.
A just-published study of dust events originating from Iceland indicated that the island is a major source of dust for the North Atlantic. This dust export can provide iron to augment phytoplankton growth. The source of the dust is the glacial outwash plains from the many ice fields of Iceland. According to the paper, retreat of the glaciers would be expected to lead to an increase of dust emissions.
The dust was monitored by a station on the island of Heimaey, just south of the main island of Iceland (and also the site of the famous Eldfell eruption in 1973; see the U.S. Geological Survey PDF document "Man Against Volcano" ). The paper noted that the dust exports were greatest in spring and early summer. The paper includes a Moderate Resolution Imaging Spectroradiometer (MODIS) image of dust plumes from southern Iceland, observed on October 5, 2004.
Screen capture of the region selected to generate the time series:
The time series shows peaks in the values of AOT 865 occurring in the spring, as discussed in the paper. The gaps in the time series are due to the fact that data was not collected during the winter when low solar zenith angles prevented the acquisition of accurate ocean radiance data. Thus, some of the highest values are observed just when data collection commenced in the spring at this latitude.
New aerosol data products based on a reanalysis of SeaWiFS data using advanced algorithms will soon be available in Giovanni.
Reference: Joseph M. Prospero, Joanna E. Bullard, and Richard Hodgkins: High-latitude Dust Over the North Atlantic: Inputs from Icelandic Proglacial Dust Storms. Science, Volume 335, 2 March 2012, 1078-1082. [ Abstract - links to full paper, which may require personal or institutional subscription to view]
A recent news article in the scientific press described a large phytoplankton bloom off the coast of Antarctica. This late-season bloom was located off of Cape Darnley, near the Amery Ice Shelf.
The ocean color radiometry 8-day data portal was used to examine the bloom. The most recently available 8-day data period is February 18-25; the next 8-day, February 26-March 4, should be available very soon. The MODIS-Terra true color image provided by the Australian Antarctic Division was acquired on March 4.
MODIS-Terra image of phytoplankton bloom off of Cape Darnley, Antarctica, March 4, 2012. The inset image shows the location of the image on the coast of Antarctica.
Giovanni chlorophyll a concentration (chl a) image for February 18-25, 2012, of the Antarctic coast. The island of Kerguelen is at center left. Increased chl a concentrations can be see off the coast.
Giovanni chlorophyll a concentration (chl a) image for February 18-25, 2012, showing Cape Darnley and the Amery Ice Shelf. Elevated chl a concentrations off the edge of the ice shelf can be seen here. The white area just off of Cape Darnley could be the phytoplankton bloom seen in the March 4 MODIS-Terra image, and interpreted as cloud due to its high reflectivity.
Giovanni euphotic depth image for February 18-25, 2012, for the Cape Darnley - Amery Ice Shelf region. Lower values of euphotic depth indicate increased light absorption and scattering by phytoplankton.
Click on any image to see it full-size.
MODIS-Aqua chlorophyll a image of the Cape Darnley - Amery Ice Shelf region for the 8-day period February 26 - March 4, 2012. The highest concentrations of chl a are seen around Drygalski Island, at approximately 65 degrees South, 93 degrees East.
MODIS-Aqua absorption coefficient for phytoplankton, February 26-March 4, 2012.
MODIS-Aqua normalized fluorescence line height (nFLH), February 26-March 4, 2012. Heightened nFLH is clearly visible near Drygalski Island.
MODIS-Aqua particulate organic carbon, February 26 - March 4, 2012.
MODIS-Aqua euphotic depth, February 26-March 4, 2012.
Using the same data products as shown above, it is clear that there is a lot of growing phytoplankton in this area. As is noted in the discussion, this is a time when sea ice should be forming, not breaking up and overturning, and it's very late in the growing season (the length of daylight and the sun angle are both decreasing rapidly), so this high level of observed phytoplankton activity seems unusual. But scientists enjoy figuring out the unusual, so these observations appear to be quite intriguing to the oceanographic community.
MODIS-Aqua normalized fluorescence line height (nFLH) image of the Princess Astrid coast of Antarctica, February 26 - March 4, 2012. The large areas of red indicate active phytoplankton along the sea ice edge, some of which may be attached to the sea ice.
MODIS-Aqua particulate organic carbon (POC) image of the Princess Astrid coast of Antarctica, February 26 - March 4, 2012. Sea ice phytoplankton communities are among the most productive in the entire world ocean.
This full-scale image from MODIS shows where the smoke was drifting over New Mexico on May 23.
During June 2012, almost exactly a year since large wildfires occurred in New Mexico (see images above), massive wildfires broke out along the Front Range of the Rocky Mountains in Colorado. On the day these images are being added to the page, the Waldo Canyon fire has raged down the mountainside in Colorado Springs, burning numerous homes, and a new fire has broken out near Boulder, potentially endangering the NOAA National Center for Atmospheric Research (NCAR).
Nitrogen dioxide (NO2) occurs in conjunction with fires, produced by oxidation of atmospheric nitrogen during combustion. The 10-day averages of the Ozone Measuring Instrument (OMI) NO2 data products for the period June 15-24 were created using the Giovanni OMI L2G (Level 2 Gridded) data portal. The images show elevated NO2 concentrations from two major fires on the Front Range, the High Park fire to the north and the Waldo Canyon fire in central Colorado. To the southwest, higher NO2 concentrations are seen from the Weber and Little Sand fires near Durango. South of the border in New Mexico, elevated NO2 is likely from the coal-fired Four Corners electricity generating station.
The Earth Observatory Image of the Day for June 26, " Wildfires Across Colorado ", shows smoke plumes from the fires observed by the Moderate Resolution Imaging Spectroradiometer (MODIS) on the Aqua satellite.