News aggregator

The Carbon Vault

Geopoetry - Fri, 06/10/2016 - 14:41
 K. Allen, 2010

Basaltic rock, Iceland. Photo: K. Allen, 2010

 

The skin of the Earth is the color of tar,

Ridged, freshly healed like the seams of a scar.

Through salt-spattered sky, a gray-winged gull sails;

Steam gently rises, the island exhales.

 

A power plant rests on porous basalt,

In spaces beneath, a dark final vault.

Carbon is cached with a strong crystal lock,

Ashes to ashes, rock back to rock.

 

______________________________________________________

Further reading:

In a First, Iceland Power Plant Turns Carbon Emissions to Stone, K. Krajick, Lamont-Doherty Earth Observatory

Rapid carbon mineralization for permanent disposal of anthropogenic carbon dioxide emissions, Matter et al., Science

Scientists Turn Carbon Dioxide Emissions into Stone, Magill, Climate Central

This is one in a series of posts by Katherine Allen, a researcher in geochemistry and paleoclimate at the School of Earth & Climate Sciences at the University of Maine.

Save

Warmer Arctic, Melting Glaciers Accelerating Greenland Ice Loss - CBC

Featured News - Fri, 06/10/2016 - 12:00
2015 was a record year for high temperatures and melting glaciers in western Greenland, an effect that is amplifying itself and could lead to accelerated warming in the Arctic, new research from Lamont's Marco Tedesco explains.

A New Solution to Carbon Pollution? - Science

Featured News - Thu, 06/09/2016 - 18:03
Researchers working in Iceland, including Lamont's Martin Stute, say they have discovered a new way to trap the greenhouse gas carbon dioxide deep underground by changing it into rock.

Martin Stute: Putting CO2 Away for Good by Turning It to Stone - The Conversation

Featured News - Thu, 06/09/2016 - 17:00
Lamont's Martin Stute writes about the CarbFix project in Iceland, where he has been working with other scientists and engineers to capture CO2 emissions and create permanent storage by turning CO2 to stone.

Iceland Carbon Dioxide Storage Project Locks Away Gas, and Fast - New York Times

Featured News - Thu, 06/09/2016 - 16:20
Lamont scientists have come up with a way to store carbon dioxide that dissolves the gas with water and pumps the resulting mixture — soda water, essentially — down into certain kinds of rocks, where the CO2 reacts with the rock to form a mineral called calcite. By turning the gas into stone, scientists can lock it away permanently.

Weird Jet Stream Behavior Could Be Making Greenland's Melting Even Worse - Washington Post

Featured News - Thu, 06/09/2016 - 16:00
Reanalyzing Greenland's last melt season, Lamont's Marco Tedesco found something odd and worrying. Greenland had shown much more unusual melting in its colder northern stretches than in the warmer south, and that this had occurred because of very strange behavior in the atmosphere above it.

Is Wacky Weather Helping Melt Greenland? - Science

Featured News - Thu, 06/09/2016 - 15:24
A new analysis of the Greenland Ice Sheet led by Lamont's Marco Tedesco points to an underappreciated culprit that could accelerate the melting of the Greenland ice sheet: wind.

Iceland Carbon Capture Project Quickly Converts Carbon Dioxide Into Stone - Smithsonian Magazine

Featured News - Thu, 06/09/2016 - 14:27
A pilot in Iceland project that sought to demonstrate that carbon dioxide emissions could be locked up by turning them into rock appears to be a success. Smithsonian Magazine talked with Lamont's Juerg Matter, who has been involved in the project, and Dave Goldberg.

Climate Change Could Force Huge Migrations Near the Equator - Washington Post

Featured News - Thu, 06/09/2016 - 12:00
New research from Lamont's Adam Sobel and alumnus Solomon Hsiang suggests that even a moderate amount of warming could force populations in the tropics to undergo huge migrations — longer journeys than they would have to take if they lived anywhere else on the planet.

Study Links Greenland Melting with 'Arctic Amplification' - UPI

Featured News - Thu, 06/09/2016 - 10:43
New research led by Lamont's Marco Tedesco links Greenland's 2015 record temperatures and melting with the phenomenon known as Arctic amplification.

Arctic's Melting Ice Creates Vicious Warming Circle - USA Today

Featured News - Thu, 06/09/2016 - 05:58
As Arctic sea ice hit a record low, scientists led by Lamont's Marco Tedesco announced the first link between melting ice in Greenland and a phenomenon known as Arctic amplification, the faster warming of the Arctic compared to the rest of the Northern Hemisphere.

Photo Essay: Seeking Humanity’s Roots

East Africa’s rift valley is considered by many to be the cradle of humanity. In the Turkana region of northwest Kenya, researchers Christopher Lepre and Tanzhuo Liu of Columbia University’s Lamont-Doherty Earth Observatory are cooperating with colleagues to study questions of human evolution, from the creation of the earliest stone tools to climate swings that have affected developing civilizations. Startling new discoveries are coming from this region at a rapid pace. Here are images from a recent field expedition. READ THE FULL SCIENTIFIC STORY or WATCH A VIDEO

 " data-cycle-speed="750" data-cycle-center-horz="true" data-cycle-caption="#gslideshow_captions" data-cycle-caption-template="{{alt}}">
Northwest Kenya's Turkwel River runs through ancient lands where many key fossils and artifacts left by early humans have come to light. The region is now inhabited by the Turkana people. Near sunset, three boys frolic on their way to fetch water. Further south, at a site called Olorgesalie, professional fossil hunter Bonface Kimeu visits some of thousands of stone axes left by proto-humans who lived here 500,000 to 1 million years ago. Bonface's father, Kamoya Kimeu, has found many of the world's most important fossils; these include a 1.6 million-year-old Homo erectus skeleton dubbed Turkana Boy, still the most complete early human remains ever found.  The Turkana region once had a climate hospitable for humans and their ancestors; now it is brutally hot and dry. Here, a local man crosses what was a lakebed as little as 5,000 years ago. The present-day Turkana people survive mainly by herding goats and camels. Near the Turkwel River, Lamont-Doherty geologist Christopher Lepre surveys layers of rocks and sediments dating back about 3.6 million years. In many places, wind, water and tectonic forces have shredded the surface into badlands, exposing many sequences across time--ideal for finding fossils and artifacts. "You get the whole human story in this one small place," says Lepre. Lamont-Doherty geochemist Tanzhuo Liu inspects eroding sediments left by a long-gone lakebed. Off near the horizon, what was probably once an island where thousands of years ago people erected stone monoliths and harpooned fish. Most research in the remote region is based out of the Turkana Basin Institute, which provides facilities for scientists. A wall at the institute documents the stream of major discoveries coming out of here. Until 2015, the stone tool displayed at lower left was considered the world's oldest, at 2.6 million years. (Actually it is a replica from a set.) Then Lepre and colleagues at Stony Brook University dated a set of tools from the Turkana region to 3.3 million years. This has reset the entire archaeological record. Lepre prepares to sample an outcrop of ancient sediments. He establishes the ages of objects found within or near layers using magnetostratigraphy--the study of how earth's magnetic field periodically reverses itself. Changes in polarity can be identified by the orientation of mineral grains. A whitish layer of compacted volcanic ash, or tuff, runs through one section. As Africa slowly tears apart along the seam of the rift valley, volcanic eruptions are commonly generated. A tuff sample chiseled out for later lab analysis. Lepre has etched an arrow into the sample to indicate which way pointed north before he removed it. Earthquakes are also frequent, as evidenced by this erosion-resistant old fault snaking though the landscape. Faults tend to confuse the chronological picture, because they thrust up some sections of land while dropping others down, putting layers out of sequence. Lepre scrambles to a high point to survey the area. Getting reliable dates means taking dozens of samples, from top to bottom. Measuring radioactive decay of certain minerals is another way to date rocks. Back at the research station, geologist Andrew Gleadow of Australia's University of Melbourne pans out heavy grains used for such analyses. A model skeleton of our own species, Homo sapiens, is used at the research station for teaching comparative anatomy to visiting students. Liu is charting past lake levels using desert varnishes--thin coatings on rocks whose compositions reflect moisture levels at the time they were deposited. Here, he hunts specimens that may have lain undisturbed for millennia. This rock has a nice varnish coating, "Look at this specimen--it's beautiful!" says Liu. Remains of the past are everywhere. By chance, the researchers spot a projectile point--maker and age unknown. They are careful to leave it in place; such artifacts are the province of archaeologists. The petrified roots of a onetime forest, from wetter times during the Pliocene, which started 5.3 million years ago. The soil and trees that grew above them are long gone, eroded away by time. Fossilized animal bones--possibly ancestors of modern antelopes or pigs--bleed from a seasonal riverbed. Nearby, in the 1990s, anthropologists found bone fragments of a distant human relative who lived around 3.5 million years ago, and may have hunted such animals. A Turkana herder stops by. Many Turkana remain deeply traditional, retaining their own language, religion and clothing. Turkana teens each carry the traditional curved stick and portable stool used for long hours of goat herding in the bush. But Kenya is modernizing fast; they also have a cell phone. The diverse hustle of Nairobi, Kenya's capital, is drawing people from many rural areas. Rapid urbanization is a global phenomenon; for the first time ever, more than half of people live in cities--a momentous step in human migration. At Nairobi's Kenya National Museum, two Homo sapiens visit the Homo erectus known as Turkana Boy. Continuing discoveries in evolution serve as powerful reminders that all humans come from, and remain, one family.
<
>
Further south, at a site called Olorgesalie, professional fossil hunter Bonface Kimeu visits some of thousands of stone axes left by proto-humans who lived here 500,000 to 1 million years ago. Bonface's father, Kamoya Kimeu, has found many of the world's most important fossils; these include a 1.6 million-year-old Homo erectus skeleton dubbed Turkana Boy, still the most complete early human remains ever found.

The See-Through Sea - The Economist

Featured News - Tue, 06/07/2016 - 12:00
The pioneering maps put together by Lamont's Marie Tharp and Bruce Heezen in the 1950s and 1960s, which first identified the structure of the mid-Atlantic ridge, were mind-expandingly right in their synoptic vision. The Economist looks at the challenges then and now of mapping the sea floor.

Was There an Ice Age in the Southern Hemisphere? - New York Times

Featured News - Mon, 06/06/2016 - 12:00
Lamont's Joerg Schaefer answers a reader's science question for the New York Times: Was there an ice age in the Southern Hemisphere?

Measuring Ice Behavior From Earth’s Glaciers to Saturn’s Frigid Moons - Science Explorer

Featured News - Thu, 06/02/2016 - 12:00
A team of Lamont researchers led by Christine McCarthy has built a new apparatus in the Rock Mechanics Lab to gain insight into the behavior of ice on Earth and elsewhere in the solar system.

Seeing the Seafloor in High Definition: Modern Mapping Offers Increasing Clarity - Earth Magazine

Featured News - Tue, 05/31/2016 - 12:00
Earth Magazine talks with Suzanne Carbotte and other scientists about advances in the mapping of the seafloor that are providing extraordinary detail.

'Dirty Blizzard' Carried Deepwater Horizon Contaminants to Seafloor - E&E

Featured News - Tue, 05/31/2016 - 12:00
Scientists led by Lamont's Beizhan Yan have discovered the mechanism that transported contaminants from the Deepwater Horizon oil spill to the bottom of the Gulf of Mexico.

Deepwater Horizon: Oil Fell on the Seabed Like Snow - Der Spiegel

Featured News - Tue, 05/31/2016 - 12:00
Researchers led by Lamont's Beizhan Yan estimate that 10 to 15 percent of the oil released by the Deepwater Horizon disaster sank to the seabed in the Gulf of Mexico, where it covered hundreds of square miles. (In German)

New Tool Measures the Behavior of Ice on Moons - American Institute of Physics

Featured News - Tue, 05/31/2016 - 12:00
A new study led by Lamont's Christine McCarthy offers a glimpse of what happens inside ice. The scientists developed a device to measure ice as it changes in response to external forces, both on Earth and on the moons of other planets.

The Database Dilemma

Chasing Microbes in Antarctica - Wed, 05/25/2016 - 12:31

Microbial ecologists know they have a problem with data archiving, particularly when it comes to sequence data.  I’m not entirely sure why this is the case; in theory it would be pretty straightforward to build a database searchable by the parameters ecologists are interested in.  I suspect that part of the problem is a lack of interest on the part of NSF (the primary source of funding for ecology) in designing and maintaining databases, and the perception that this is duplicative of the (rather sad) efforts of the NIH-funded NCBI in this area.  I don’t want to be too disparaging of NCBI – they have a much harder job than the simple database I’ve outlined above – but the Sequence Read Archive (SRA) and other NCBI repositories are a near-total disaster.  We’re wasting a lot of collective time, money, and effort developing sequence data that can’t be used for future studies because it can’t be easily accessed.

In response to this a small number of alternate repositories have popped up.  In my experience however, they don’t offer much of an improvement.  This isn’t meant as a harsh criticism of various peoples’ good efforts, I just don’t think the community’s found a viable model yet.  One example is the VAMPS database operated by the Marine Biological Laboratory, which in addition to having an unfortunate name (I believe it was named before the onset of the angsty-teen vampire thing, try googling “VAMPS” and you’ll see what I mean), tries to do too much.  Rather than organizing data and making it searchable the designers opted to try and build an online analysis pipeline.  Instead of easy to find data you end up with an analytical black box and canned results.

The reason for all this soap-boxing is my experience this week trying to obtain some global marine 16S rRNA gene amplicon data for a project that I’m working on.  The data I had hoped to obtain was used in the 2014 PNAS paper Marine bacteria exhibit a bipolar distribution.  My difficulty in obtaining the data highlights failures at the author, reviewer, journal, and repository levels.  I have great respect for many of the authors on this paper, so again this isn’t meant as a harsh criticism (I’m sure I haven’t made data I’ve published sufficiently accessible either), but rather a lesson in something that we all need to do better at.  I’ll end with the Python-based solution I came up with for getting useful sequence data out of the VAMPS flat-file format.

The 2014 paper used 277 samples that were collected as part of the International Census of Marine Microbes (ICOMM).  There are quite a few things one can say about that particular endeavor but we’ll save it for another post.  The authors included the accession numbers for all these studies in a supplementary file.  This file is a pdf, so not the easiest thing to parse, but at least the text is renderable, so the (7-page) table can be copied into a text file and parsed further.  With a little bit of effort you can get a single-column text file containing the accession numbers for each of these studies.  Great.  Unfortunately a good number of these are wildly incorrect, and most of the remainder point to SRA objects that can’t be recognized by the woefully inadequately documented SRA toolkit.  Here are some examples:

SRX011052 – Enter this accession number into the SRA search bar and it will navigate you to a page for a sample with the helpful name “Generic sample from marine metagenome”.  Nevermind for a moment that “metagenome” is an inappropriate name for an amplicon dataset, where is the data?  Click the link for “Run” (with a different accession number: SRR027219) and navigate to a new page with lots of obscure information you don’t care about.  If you’re actually doing this you might, in near desperation, click the “Download” tab in the hope that this gets you somewhere.  Psych!  All that page has is a message that you need to use the SRA toolkit to download this data.  Let’s try the prefetch utility in the toolkit:

>prefetch SRR027219 2016-05-25T15:30:43 prefetch.2.4.2 err: path not found while resolving tree within virtual file system module - 'SRR027219.sra' cannot be found.

Whatever that means.  There’s a trick to using this utility, which I’ll post as soon as I remember what it is.  The documentation suggests that the above command should work.  The absurdity of needing to rely on this utility in place of FTP usually drives me to using the EMBL-EBI site instead.  This is the Euro equivalent of Genbank, anything submitted to one should be accessible through the other.  In general EMBL-EBI is much better organized and more user friendly than SRA.

Entering “SRX011052” into the ENBL-EBI search bar returns some results that also point us to SRR027219.  The page for SRR027219 has such helpful things as an FTP link for direct download as well as some basic info on the sample.  That’s great, but because the authors opted to provide “Experiment” accession numbers for some samples (e.g. SRX011052) there is no way that I can see to generate a script to automate downloads, as we have no way of knowing the run accession numbers.  The solution is to tediously follow the links to all 277 runs.  I actually started doing this which lead me to my second problem.

SRP001129 – At some point in the supplementary table the accession numbers switch from a “SRX” prefix to “SRP”.  Picking one of these at random and entering it into the EMBL-EBI search bar returns a result for a genome sequencing project related to the Human Microbiome Project.  Clearly this isn’t from the ICOMM project.  Roughly a third of the accession numbers reported in the paper aren’t correct!

Clearly I wasn’t going to get the data that I needed through either EMBL-EBI or the SRA.  Knowing that the study had been part of the ICOMM I turned to the MBL VAMPS database where the ICOMM data is also archived.  Thinking that I was on the home stretch I followed the link to the ICOMM portal from the VAMPS homepage.  The right side column gave me two options for download; tar-zipped sff files or a tab-delimited file.  Not wanting to deal with converting from sff (the organic 454 file format) I took a look at the tab-delim option.  The top lines of the file look like this:

project dataset sequence taxonomy knt rank total_knt frequency ICM_ABR_Av6 ABR_0006_2005_01_07 GAAACTCACCAAGGCCGACTGTTAGACGAAGATCAGCGTAATGAGCTTATCGGATTTTCAGAGAG Archaea;Euryarchaeota;Thermoplasmata;Thermoplasmatales;Marine_Group_II 1 family 36172 2.7645693e-05 ICM_ABR_Av6 ABR_0006_2005_01_07 GAAACTCACCAAGGCCGACTGTTAGATGAAGATCAGCGTAATGAGCTTATCGGATTTTCAGA Archaea;Euryarchaeota;Thermoplasmata;Thermoplasmatales;Marine_Group_II 1 family 36172 2.7645693e-05 ICM_ABR_Av6 ABR_0006_2005_01_07 GAAACTCACCAAGGCCGACTGTTAGATGAAGATCAGCGTAATGAGCTTATCGGATTTTCAGAGAG Archaea;Euryarchaeota;Thermoplasmata;Thermoplasmatales;Marine_Group_II 13 family 36172 0.000359394006 ICM_ABR_Av6 ABR_0006_2005_01_07 GAAACTCACCAAGGCCGACTGTTAGATGAAGATCAGCGTAATGAGCTTATCGGATTTTCTAGAGAG Archaea;Euryarchaeota;Thermoplasmata;Thermoplasmatales;Marine_Group_II 1 family 36172 2.7645693e-05 ICM_ABR_Av6 ABR_0006_2005_01_07 GAAACTCACCAAGGCCGACTGTTAGATGAAGATTCAGCGTAATGAGCTTATCGGATTTTCAGAGAG Archaea;Euryarchaeota;Thermoplasmata;Thermoplasmatales;Marine_Group_II 1 family 36172 2.7645693e-05 ICM_ABR_Av6 ABR_0006_2005_01_07 GAAACTCACCAGGGCCGACTGTTAGATGAAGACCAGTGTAACGAACTTGTCGGATTTTCAGAGAG Archaea;Euryarchaeota;Thermoplasmata;Thermoplasmatales;Marine_Group_II 1 family 36172 2.7645693e-05 ICM_ABR_Av6 ABR_0006_2005_01_07 GAAACTCACCAGGGCCGACTGTTATATGAAGACCAATGTGATGAACTTGTCGGATTTTCAGAGAG Archaea;Euryarchaeota;Thermoplasmata;Thermoplasmatales;Marine_Group_II 2 family 36172 5.5291386e-05 ICM_ABR_Av6 ABR_0006_2005_01_07 GAAACTCACCAGGGCCGACTGTTATATGAAGACCAGCGTAATGAGCTTGTCGGATTTTCAGAGAG Archaea;Euryarchaeota;Thermoplasmata;Thermoplasmatales;Marine_Group_II 1 family 36172 2.7645693e-05 ICM_ABR_Av6 ABR_0006_2005_01_07 GAAACTCACCAGGGCCGACTGTTATATGAAGACCAGCGTGATGAACTTGTCGGATTTTCAGAGAG Archaea;Euryarchaeota;Thermoplasmata;Thermoplasmatales;Marine_Group_II 1 family 36172 2.7645693e-05

Okay, so this is a bizarre way to share sequence data but at this point I was willing to take it.  Basically the file gives us unique reads, their abundance (knt), and taxonomy according to VAMPS.  Want I needed to do was convert this into a separate, non-unique fasta file for each sample.  With a little help from Pandas this wasn’t too bad.  The vamps_icomm_surface.csv file specified below is just a csv version of the pdf table in the PNAS publication, after some tidying up.  The second row of the csv file is the sample name in VAMPS (I had to split the first column of the original table on ‘.’ to get this).

import pandas as pd ## get the metadata for surface samples, this was generated ## from the pdf table in the PNAS paper metadata = pd.read_csv('vamps_icomm_surface.csv', index_col = 1) ## get the seqs, downloaded from https://vamps.mbl.edu/portals/icomm/data_exports/icomm_data.tar.gz seqs = pd.read_csv('icomm_data.tsv', delimiter = '\t', index_col = 1) ## iterate across each study in metadata, making a temp dataframe ## of reads from that study for study in metadata.index: temp = seqs.loc[study] temp_bacteria = temp[temp['taxonomy'].str.contains('Bacteria')] ## now iterate across each row in the temp dataframe, exporting ## the sequence the specified number of times to a fasta file with open(study + '.bacteria.fasta', 'w') as output: for row in temp.iterrows(): sequence = row[1]['sequence'] for i in range(1, row[1]['knt']): name = row[0] + '.' + str(i) print name print >> output, '>' + name print >> output, sequence

To end I’d like to circle back to the original dual problems of poor database design and erroneous accession numbers.  Although these problems were not directly connected in this case they exacerbated each other, magnifying the overall problem of data (non)reuse.  Some suggestions:

  1.  The SRA and other databases needs to aggregate the run accession numbers for each new study.  A couple of clicks should lead me to a download link for every sequence associated with a study, whether that study generated the raw data or only used it.  I don’t believe it is currently possible to do this within SRA; initiating a new project assumes that you will contribute new raw sequences (correct me if you think otherwise).  This should be easy to implement, and the association of a publication with a project should be rigorously enforced by journals, departments, and funding agencies.
  2. Accession numbers should be automatically checked by journals the way references are.  If it points to something that isn’t downloadable (or at least has a download link on the page) the author should correct it.  There is no excuse for incorrect, incomplete, or non-run accession numbers.  That this was allowed to happen in PNAS – a top-tier journal – is double frustrating.  And frankly, if you reviewed this paper you owe me beer for not spot checking the refs.
  3. Validating this information is the job of the reviewers.  Almost nobody does this.  Usually when I review a paper containing sequence data I’m the only reviewer that insists the data is available at the time of publication (I can’t think of a single case where I wasn’t the only reviewer to insist).  Not only is this a necessary check, but part of the job of the reviewer is to spot check the analysis itself.  Clearly, without data that’s not being done either…

And that’s enough ranting about data availability for one day, back to work…

Pages

 

Subscribe to Lamont-Doherty Earth Observatory aggregator