Citizen Science at AGU 2015 Fall Meeting

I went to my first meeting of the American Geophysical Union in December. It was quite the experience; I’ve never seen academic conferencing on such a scale before. I liked the primarily-posters format because it was much more interactive overall–I could linger and discuss where I wanted to and skip the stuff that was less interesting to me. And to my surprise, there was a lot more that interested me (mostly in the earth and space informatics section) than I had initially expected.

However, it was hard to find the citizen science content, aside from that which was labeled as “education” despite a primary focus on science over outreach. With such a massive program, it’s pretty important to be able to search effectively, and I missed a lot of good stuff just because I didn’t know how or where to find or look for it. I made it to just 2 oral presentation sessions that featured citizen science in the session title; most other citizen science presentations were unobtrusively tucked into sessions with titles that presumably focused more on the science than the process and participants.


 

Climate Literacy: Improving Climate Literacy through Informal Learning & Citizen Science 1
December 12, 2015

Realizing the Value of Citizen Science Data, Waleed Abdalati

Perspective matters: diversity of the public is part of benefits. He was NASA Chief Scientist at time of Curiosity landing.

4 part series of TV segments on citizen science, starting with CBC at Everglades. Then gives the example of NPN & Nature’s Notebook. Points out that these are good data because people really care, as much or more than professional scientists. CoCoRaHS is another example, video of NWS staff setting up an alert based on CoCoRaHS data, process between report and radio alert is 2-3 minutes.

Another series – the crowd & the cloud. #1 Even big data starts small; #2 Viral vs. virus; #3 Feet in the field, eyes in the sky; #4 Citizens 4 Earth. Smartfin – surfer science.

Fantastic high quality video, really compelling teaser for the series. Will air in 2017 on PBS.

Q: this takes skill, how is training done?
A: CoCoRaHS has training protocol.


Citizen Science Contributions: Local-scale Resource Management and National-scale Data Products, Jake Weltzin

“from kilometers to continents”

Monitoring for decision-making at Valle de Oro NWR, first urban wildlife refuse in Albuquerque. Needed decision info and also public engagement. Data presented with bar chart that shows when migratory species are present at refuge so they know when to manage for them. Also working on wetland restoration–reducing Siberian elm and increase Rio Grande cottonwoods. Checking the flowering an fruiting–Siberian elm is leafing and flowering about a month ahead of cottonwood, and you need bare ground for cottonwood to propagate, so they need to remove the Siberian elm in the month before cottonwoods in order to prompt cottonwood growth.

Product framework: phenology status data goes to phenometrics; climate and meteorological data goes to phenology models and integrated data sets; remote sensing data also goes into integrated data sets; phenometrics goes into phenology models; final products are gridded maps and datasets, and short-term forecasts.

Showing NCA annual start of spring based on lilac data. Very pretty maps of “PRISM data set” for start of spring, 4km scale national map. Local version maintains granularity and scales down to NPS locations, so you can see first leaf index for park locations. But NPS cares less about when things happen than change from historic record, since the data go back to 1900, they can show biological response to climate change at level of national parks.


Putting citizen-collected observations to work — CoCoRaHS, Nolan Doesken

Starts with funny 2-minute animated intro “each measurement is like a pixel in a picture”. Talks about 1997 flood in Fort Collins–60% of library holdings were destroyed as they were in the basement due to work on upper floors. Recent expansion into Canada and Bahamas; now has over 20K volunteers.

Goals are quality precipitation, and also education & outreach. Easy low-cost equipment is important–gauge is equivalent to that used for historic climate monitoring that NOAA does, therefore can fit into long history of measurements. Mobile app for data submission as well as web forms; permanently archive data and provide raw data and summary reports. Data are good for supplementing other sources like COOP.

Data tend to be accurate, spatially detailed (except in Nevada–not enough people), timely, etc. Who uses data? Weather forecasters, hydrologists, water management, researchers, agriculture, climatologists, health, insurance industry, tourism. Data are fed into weekly US Drought monitor process, drought conditions are reducing. Snow data is hard to get so their sources are valuable. National Hurricane Center uses the data in post-storm summaries to describe the impacts.

Challenges: owl sitting on top of rain gauge! Much more male than female, very white, mostly college educated. Age demographics leans toward older, those who stick with it tend to be from that demographic even though they have good rates of signup for younger and more diverse demographics. Recruiting at national scale is tough. They have over 250 local volunteer leaders; need to recruit and train 3K new volunteers per year to balance attrition.

Cost effective but not free; after 18 years, still hanging on. Photos of a bear checking a rain gauge.

Q: GEO group looking for improving in situ precipitation measurements, especially in Africa. How to export to Africa?
A: It’s a matter of finding local leaders who care about local precip. Putting local face on project is more compelling than most other options. Subsidize the rain gauge cost, and then communication is the next consideration–need infrastructure.


Crowdsourcing science to promote human health: new tools to promote sampling of mosquito populations by citizen scientists, Rebecca Boger

GLOBE program–international citizen science in the classroom, 20 years old. Discussing how materials are developed, new mosquito larvae protocol.

Train-the-trainer model with F2F workshops–big backlog and long waits to join program, so moving into an LMS. Developing training slides for 50+ protocols, will be available in 2016, emphasis on knowing how to conduct protocol and not pedagogy. They have to pass quizzes before being able to set up a login and get full access.

Developing a mosquito monitoring protocol: can do genera ID with hand lens, species ID with microscope and experts. Sampling from containers as well as ponds, streams, puddles. Lots of research questions students can explore with the data. Have to get it up at the end of the year; will be doing a field campaign early next year to launch new protocol.


Era of Citizen Science and Big Data: Intersection of Outreach, Crowd-Sourced Data, and Scientific Research 1
December 18, 2015

The Citizen CATE Experiment for the 2017 Total Solar Eclipse, Matthew Penn

Working with 3 government research labs, 3 corporate partners, 4 universities, 3 K12 teachers, and participants. Donating telescopes to observers after event, sponsors include the companies who make software and filters.

Upcoming eclipse on August 21, 2017 will drive tourism, will be most viewed eclipse in history. Total eclipse opens up a window for viewing the inner corona in a way we can’t from space. The part easily viewed from an eclipse is the hardest part to study from a spacecraft. Planning to look at what is happening with polar plumes–they’re interesting but they need more data than observations from 3.5 minutes from one location. Looking at the eclipse in Mongolia in 2009, they knew they would be able to see scientifically interesting events.

Path of totality goes from PNW to South Carolina, to plan is to provide identical telescopes or volunteers to use at specific locations, transfer ownership after the event, and support ongoing use of telescopes. While it will be viewed for only 2.5 minutes in each single location, the entire path of totality is 90 minutes.

Funding needs about $180K for equipment alone: 60 sets of telescopes, filters, software, mount and drive, still need $ to cover cameras and laptops. Expecting about 26 GB data per site, 1560 GB (TB?) in total. Sending data via 3-day priority mail, equivalent of 6 MB/second, upload of about 2 GB on day of eclipse itself.

Afterward, they’re looking to develop additional projects for work on comets (can’t get major telescope time), solar programs on sunspots, and variable stars with prototype equipment in partnership with AAVSO.

Proof of concept: Did one day of training with a volunteer who was going to Faroe Islands in Mar 2015, conditions were lousy, but for 15 seconds got data of inner corona. The harder job was shipping it around the world and using crummy software. For a more prepared test, doing a train-the trainer training with 5 locations in Indonesia for March 2016 to verify process.

Interested? mpenn@nso.edu; mpenn@noao.edu; sites.google.com/site/citizencateexperiment

Q: how much does weather matter? A: weather isn’t great for about half the range, tends to be 60% cloudy. But with 60 sites they should get good coverage, and they’re hoping for 100, but if they add more they have to add where it’s cloudy to hopefully get more data from the sparser areas. Some range of +/- 10 miles to move in, but expect some gaps.


Synergetic Use of Crowdsourcing for Environmental Science Research, Applications & Education, Udaysankar Nair

Motivated by needs for data that aren’t collected by agencies but suited to crowdsourcing with compute platforms like Google Earth Engine.

Using ODK, “end to end design” of system, that pushes data to Fusion Tables and Google Earth Engine, merged with sat imagery from NASA via a maps engine.

Land Use & Land Cover Change data currently relies on remote sensing data, but it needs ground truthing for contextual information. Many potential uses for data.

Claims 4m accuracy for GPS on app. Can use ODK offline to collect data–step by step overly simplified form, usability could be problematic.

Tested with a MS classroom, introducing with the topic of biomes. Requires lesson plan, including learning standards. Had kids use mobiles with ODK to track land cover in their neighborhood. Also did some work with student teachers in India for mapping small water bodies to support Kerala State Biodiversity Board. Also looking at collecting data on open water containers for vector borne disease research; frost occurrence; damage after severe weather. Doesn’t mention how this is fed back to students.


 

LastQuake: Comprehensive Strategy for Rapid Engagement of Global Earthquake Eyewitnesses, Massive Crowdsourcing, & Risk Reduction, Remy Bossu

Points to eBird: you can’t do this for earthquakes because target reporters are eyewitnesses. Focus on felt earthquakes, looking at SM activity and speed of feedback so info needs to be available across SM platforms. QuakeBot, apps & add-ons are intended to automatically merge direct & indirect eyewitness contributions, seismic data, and other sources.

Can’t identify “felt EQ” with instruments, but can via SM. Just look at tweets with earthquake in them in US. But not every place uses Twitter that much. They use real-time web traffic to their authoritative site to figure this out based on IP addresses, could tell Kathmandu had not been flattened b/c they continued getting visits after several minutes.

citizenseismology.eu, @LastQuake

During Nepal event, made automatic map but did not predict intensity until about 19 minutes, confirmed damage at 20 minutes, published 38 tweets in that time during which there were main shock and 5 felt aftershocks. Working to develop an app, UI improvements, gets better geolocated pics & videos, sharing comments to SM, and push notifications. Got decent data despite the fact that the quakes they have recorded were areas where LastQuake aren’t well known. They validate pics for lack of IP infringement, respect for human dignity, and accuracy to known issues.

Quick rise in app downloads 10 min after Nepal. After 9 days, had identified most of quakes post-main-event. 85% of access from Nepal via Mobile, with 1/3 via app & 70% of reports. Traffic picks up in under a minute of shocks. Case on December 7: 110K downloads, 82K in operation (75% retention). Saw app launched within 1 minute of event and notifications: immediate response worldwide.

One KPI is number of responses within 30 minutes. Examples where they aren’t well known: Afghanistan, Arizona, England, Malaysia–hundreds of responses in each case, 2400 for AZ.

How were people finding them? This is only app providing info on felt earthquakes. It only takes hours for info to be shared. So they asked for feedback–what improvements? They wanted help, what to do in earthquakes. So developing visual pop-ups with do’s and don’t’s: visual popups (stay away from buildings, don’t call 911 unless injured), adding an “I am safe” button. Seeing this as risk reduction information for public: reduce inappropriate behaviors and fatal errors.

Q: Implications of using this for catastrophes, like anthropogenic disasters like shooters? How can you verify the truthing, veracity of content?

A: Rapid-onset events are easy to tell: eyewitnesses hit the website within 2 minutes and others don’t know of the event yet to falsify, but floods are much harder to tell. They don’t see a lot of people messing with the comments, the “not so bright” people are easy to spot. Can easily remove outliers, likely not because they are lying but because they are so emotional. With pictures, it’s not about reliability–photo of small crack in the wall isn’t useful to them, care more about larger damage.


CosmoQuest: Building community around citizen science collaboration, Pamela Gay

Data landscape for space science data is changing dramatically–“horrific data flood flying down upon our heads and across our internet connections”. Need help handling tons of data. Can’t get enough postdocs, have to open the doors to the ivory tower. Open data and open access will help, but requires supporting community: curricula, projects specific to grade level, adult learning, planetarium & science on sphere content to recruit and disseminate, crowdsourced podcasts, “guerrilla science” at science-related events.

Current projects focus on surface science. CitizenScienceBuilder for image annotation. TransientTracker for photometry and other products. Building data products and simulations. Portals like Moon Mappers, Asteroid Mappers, etc. Funded through 2020 with some pre-selected projects, but if all goes well, there will be an RFP for projects with details on how to ensure that results get published for small grants up to $60K. Providing educational materials, curricula, etc.

They partner with a lot of programs for podcasts, live YouTube events with up to 5K attendees, in-person events. “Come science with me”.


A method to harness global crowdsourced behavior to understand something about avalanches, Jordy Hendrikx

Snow avalanches cause 30 deaths/yr in US, up to 500 fatalities worldwide. $1B damage in US alone. Also dramatic uptick in backcountry users, and fatalities increase slower than usage so education is likely helping. Historically 4 parts to risk: snowpack, weather, terrain, and most importantly people. Need to understand their decisions.

Research that tries to look at causes of avalanche fatalities tries to understand accidents based on result, but fatalities are usually a cascade of errors so it’s hard to figure out a causal factor. Rather than a consequence of a series of decision, try to go to “top of cliff” to figure out which groups more likely to be at risk in future due to behavior, and then use targeted education. Goal is prevention via behavioral understanding.

Crowdsourcing by taking realtime GPS tracks on a smartphone app, then do Internet surveys about decision-making they can connect to it. Using a marketing approach to decisions. They describe & quantify travel practices in concert with group decision making dynamics and participant demographics, using GPS track as expression of decisions and terrain use.

Sending people to webpage sounds easy but is hard, need to advertise and get word out, that’s harder than scientists think it will be. Then they show simple flowchart–sign up, download app, track trips, auto-reply afterward, fill in survey. Have been doing this since 2013/4 season, noticing that there’s self-selection bias among who participates and trying to grow sample to broader range so as to get more behavioral insight. Using snowball sampling via SM, word of mouth, but have to reflect culture of a crowd, not stuffy white lab coat. Getting thousands of track from around the world.

Outreach is critical–presenting at workshops & public events, publications in popular press.

 

Looking Back, Moving Forward in PPSR

Conference on Public Participation in Scientific Research, Day 1 Session 1 – 8/4/2012

—–

PPSR: How We Got Here and Where We Go Now
Abe Miller-Rushing

Exciting to bring wide range of disciplines together and develop a more global perspective. Take-aways: PPSR is not new, has always been important to science, and is growing and innovating very quickly.

Models for PPSR – taxonomy of projects by degree of involvement of participants, contributory, collaborative, and co-created.

How we got here: Science began as amateur research with Plato and Aristotle, often by rich people who had money and time. Some of the most important science has relied on public participation, e.g. Linneas. Professionalization of science has marginalized public participation and that’s where we are today.

Nonetheless, PPSR has continued, just not always labeled as such or recognized as broadly as it should be. Originally it was often specimens, but now it’s usually observational data. It’s also used to solve local problems, and that’s an important role. Big data sets are also being generated through citizen science, some of these are the most important for their field, for example NOAA’s weather data that is being used to understand climate change.

Recent developments: huge improvements in tech, communication, data storage, analysis & best practices, this kind of revolution is not without precedent. Another big advance is with explicit participant-focused outcomes, which is still fairly new.

We need data over wide time periods and geographic ranges to achieve many of our scientific goals moving forward. Things like fine-scale weather observations through CoCoRaHS which is really important for decision-making; looking at changes in phenology to understand climate change and losses in biodiversity, e.g. findings from looking at Thoreau’s data through time with current citizen science data.

Many applications: images and sound analysis; real-time data for near-term predications; collection and transcription of historical records; health and environmental justice.

Huge growth in PPSR recently, e.g. with ISI on peer-reviewed publications – exponential growth in last 6 years. More to come as cross-disciplinary dialogue, collaboration & innovation develop. Now we see a need to formalize and support the field and practitioners. Still getting push-back at NPS for making management decisions based on PPSR data.

Where do you want PPSR to go? What does the field need? What do you need in your role in PPSR? What should an organization for PPSR do? Poster session opportunities to post your responses to important questions. Will be using this feedback in closing session discussion – please participate and help us act on these recommendations from the community.

Q: Issues – recognition of citizen science and use of the language in publications – what is this doing to help legitimize PPSR.

Grand Challenges and Big Data: Implications for PPSR
Bill Michener

Challenges we face, scientifically and technologically, focusing on data issues as that’s where the rubber hits the road from a science perspective. Many issues we are concerned about, primarily related to climate change, clean energy, and so on. We’re in a new age where we’re hitting some tipping points and likely to see very abrupt changes that will have significant impact on future and quality of life on earth.

Many tools being used for data-intensive science, but data management is one of the challenges standing in the way of results – we need to speed up time to results and reduce time on mundane tasks like data management. Another key challenge is expanding participation.

Major concern – where are the data? We need to be able to integrate data to address major scientific challenges. This leads to the long-tail distribution of data problem with many data orphans. Jim Gray – “Most of the bytes are at the high end, but most of the datasets are at the low end.” Brings up an important question – we’re all familiar with the research life cycle, but how do we link it to the data life cycle?

Solutions: DataONE is addressing some key issues, e.g. data preservation. Intro to http://www.dataone.org.

One of the main science data management/analysis tools currently in use is Excel [shifting toward Google Docs]. D1 is developing tools for R, which is one of the second-most popular tools for analysis, working with many partners.

Issue 2: Data discovery – not easily found with traditional search tools. Major project of D1 has been ONEMercury for searching across datasets.

Issue 3: Tools for innovation and discovery. We’re in the 4th paradigm of research, focus on data-intensive research that requires new tools, techniques, and ways of doing research. Another way the investigator toolkit fits in to address this question. Examples include DMPTool, data management planning tool – helps get grants funded for agencies requiring data management plans, but should be a consideration for PPSR projects. Supports 12+ templates required by different agencies with walk-through series of steps to address required points for data management plans. At the most, all you need to do with this after going through wizard is change font.

Upcoming tool: DataUp to check Excel spreadsheets for best practices, create metadata and connect to ONEShare, one of the D1 repositories – all for free.

Finally, need for tools for exploration, visualization, and analysis. Example needed data layers from several sources to address research question, one only found through word of mouth, had to develop new modeling tools and work with new tool (VisTrails) to develop visualization.

People and Participation: Educational and Community Components of PPSR Projects
Heidi Ballard

Came to PPSR as HS teacher, then worked in community-based forest management, and then working with science education at UC Davis.

Need to look at PPSR across different practices. Many disciplines and goals represented here: biochem, ecology, astronomy, nat rsc mgt, and public health. Also have: psych, sci & enviro ed, social justice & community development, sociology, anthropology.

Names for PPSR categories are called different names by other scholars, this is explored in recent Ecology & Society article. Need to think about more than degree of participation and what part of scientific process. Other important aspects related to quality – Whose interests are being served, and to what end? Who makes decisions? Who has the power?

This leads into discussions of democratizing science – role of PPSR is improved science understanding for everyone, which results in better research. Examples include LiMPETS monitoring 600 miles of CA’s National Marine Sanctuaries, students taking it very seriously because it will be used for science. Additional examples related to rice growing and health impacts.

Looking at social and educational outcomes of PPSR – individual, programmatic, & community level. Often PPSR focuses on programmatic level – audience reach, engagement, program strengths/weaknesses, etc. Current work focusing on individual learning outcomes, and community-level outcomes are exciting area to develop – social capital, community capacity, economic impacts, trust between public, scientists & managers.

Her main question is: if we think about intertwining of social & ecological systems, many stakeholders involved, can we improve their resilience?