The PEcAn project has been selected to participate in Google Summer of Code(GSoC) once again for 2018. We are excited to work with more great developers. Information about our project and how to apply can be found on our GSoC page.
The application period is from March 12th-27th, but we welcome all interested developers to contact us to discuss projects and/or the application process and start coding!
The PEcAn project will be holding a two-part event this year at the 2017 AGU Fall meeting on Tuesday December 12th, 2017 from 4-7pm. The meeting will include an open demonstration (4:00-5:30) followed by a focused science team collaborator meeting (5:30-7:00).
The Event will be held at the Hilton next to the New Orleans convention center:
Hilton New Orleans Riverside, 2 Poydras St., New Orleans, LA 70140 Third Floor - Parish Room
The demonstration will entail a general overview of the project and breakout groups in which members of the PEcAn team will lead hands-on tutorials and take questions. Attendance is free and open to whomever shows up, so please encourage your colleagues and collaborators to come.
The Science Team meeting will focus on high-level development goals and user needs for the coming year. This will be the first year we’re opening up the Science Team meeting to beyond the core development team. The meeting is open to all, but we’re particularly interested in touching base with projects that are funded to use PEcAn or are otherwise invested in setting up PEcAn-based projects. Confirmation of attendance is appreciated but not required.
The PEcAn Team will be at the 2017 American Geophysical Union (AGU) Conference and will be presenting the following talks and Posters:
ORAL PP21D-06: Istem Fer "Paleoclimate reconstruction through Bayesian data assimilation", 9:15 Room 343
ORAL Poster Hall D-F: Cristine Rollinson "Impacts of weather versus climate and driver uncertainty on multi-centennial ecosystem model simulations", 10:35 Room 343
ORAL PP22A-05: Ann Raiho " Estimating the ecology of extinct species with paleoecological data assimilation", 11:20 Room 343
ORAL B22C-07: Alexey Shiklomanov "Leaf optical properties shed light on foliar trait variability at individual to global scales ", 11:50 Room 383-385
POSTER IN31A-0063: Rob Kooper "Data Flow for the TERRA-REF project", 8:00 Poster Hall D-F
PANEL B31H-04: Michael Dietze "Integration of Ecosystem Research Infrastructures for Multiscale Analysis II", 8 - 10 am Rooms 383-385
Poster B31A-1979: Eugenie Euskirchen "Predicting Changes in Arctic Tundra Vegetation: Towards an Understanding of Plant Trait Uncertainty", 08:00 - 12:20 Poster Hall D-F
ORAL David LeBauer: "TERRA REF: Advancing phenomics with high resolution, open access sensor and genomics data", 10:40 Room 383-385
ORAL Tony Gardella: "A multi-model assessment of terrestrial biosphere model data needs", 2:25 Room 356-357
Poster B51G-1893 Anne Thomas: "Assessing Ecosystem Model Performance in Semiarid Systems", 8am Poster Hall D-F
Poster B53B-1953 Molly Aufforth "Constraining Agricultural Irrigation Surface Energy Budget Feedbacks in Atmospheric Models", 1:40 pm Poster Hall D-F
ORAL H54D-01: Betsy Cowdery: "Looking Past Primary Productivity: Benchmarking System Processes that Drive Ecosystem Level Responses in Models", 4 pm Room 298-299
ORAL H54D-05: Shawn Serbin: "Evaluation of Multiple Mechanistic Hypotheses of Leaf Photosynthesis and Stomatal Conductance against Diurnal and Seasonal Data from Two Contrasting Panamanian Tropical Forests ", 5 pm Room 298-299
ORAL B54B-07: Tess McCabe: "Integrating invasive grasses into carbon cycle projections: Cogongrass spread in southern pine forests", 5:30 pm Room 388-390
Istem attended the FIA Stakeholder Science Meeting this year in Park City, Utah. She presented her work on assimilating demographical data into model and validating model against FIA data.
Alexey and Betsy attended the Google Summer of Code Mentor Summit as a result of our project's successful participation in the Google Summer of Code Program. We look forward to participating again next year and having more developers successful contribute to our project.
Mike Dietze Attended and presented at the 2017 Cost-Profound FP1304 Final Meeting. His presentation was about his "PROFOUND thoughts on Model-Data Integration & Forecasting (pun intended)."
Mike Dietze, Ankur Desai, Betsy Cowdery, and James Simkins are all returning to teach at this year's Flux Course.
Shubham graduated with an Integrated M.Tech in Mathematics and Computing'15 from IIT Delhi. He then worked as a Data Scientist for a year (2015-16) at TrulyMadly Matchmakers Pvt. Ltd, in the online dating domain. Shubham is currently enrolled in the one year, Masters (MSIAM) Data Science specialization program offered jointly by Universite Grenoble Alpes and Grenoble INP, Grenoble, France. He will be working on our GSoC project 'Scientific Visualization' over the summer, specifically responsible for our Shiny Dashboard visualization of our model endpoints.
Aman Kumar graduated with a Bachelor of Engineering from BMS Institute of Technology & Management Bangalore, India. He has always been a tech enthusiast and web technologies, Offensive security, Distributed technologies, Big data interest him the most. Currently, he is working on replacing the current VM distributions technique with docker containers and modularizing it into smaller independent parts like PEcAn-Core, PEcAn-web, Postgresql & Bety so it will become easy to maintain.
This summer, Anne Thomas will be contributing to the benchmarking module of PEcAn, putting together multi-model benchmarks for semi-arid and arid ecosystems as well as working on the overall benchmarking workflow. A lifelong nature lover and a growing programming nerd, Anne is studying Conservation Biology and Bioinformatics at Brigham Young University. After she completes her undergraduate degree she plans to pursue a PhD in computational ecology with the goal of addressing conservation challenges like climate change with programming power.
Mike Dietze attended the 3rd PROFOUND Spring School: Baysian calibrartion, forecasting and multi-model prediction of process-base vegetation models in Rencurel/Grenoble, France as a lecturer.
A report fom the course can be found by clicking on this link.
Mike Dietze attended a Data Assimilation Workshop held by NSF in Arlington, Virginia
Three PEcAn team members participated in the American Association of Geographers (AAG) Annual Meeting in Boston this year.
A number of PEcAn team members were in attendance at the Joint North American Carbon Program (NACP) and Ameriflux Principle Investigators Meeting involved in multiple activities.
Mike Dietze, Ankur Desai and other members of the team led a break out session, giving a demo of PEcAn's capabilities while discussing the needed tools and goals for the carbon cycle prediction in our community.
Mike Dietze: Carbon cycle forecasts need to be closer to real-time: linking theory, data, models, and cyberinfrastructure
Istem Fer: Improving forest carbon sequestration predictions by constraining demographic processes in an ecosystem model
Alexey Shiklomanov:Leaf optical properties shed light on foliar trait variability at individual to global scales
Tempest McCabe: Validating the Ecosystem Demography Model Version2 for Southern Pine Forests
Anthony (Tony) Gardella: Modeling Community Cyber-Infrastructure Needs and Opportunities
Mike Dietze and Kenton HcHenry had a poster at the 1st NSF Data Infrastructure Building Blocks (DIBBS) PI Workshop : "Brown Dog - A Science Driven Data Transformation Service", highlighting BrownDog's capabilities to "access and preserve data transformation tools, track provenance, track information loss, manage data movement, and process jobs in a scalable manner across a diverse set of computational resources".
We will be holding an open house & social gathering for the PEcAn project on the Monday night of AGU
When: Mon, Dec 12th from 6:45pm -10:30pm
Where: The Garage SFO- Luxury Live/Work Warehouse. 951 Folsom St, San Francisco, CA 94107 (~2 blocks south of Moscone)
There will not be any formal presentations, so feel free to come and go as it fits your schedule.
We encourage you to:
Food and drinks will be provided.
To help us with planning, we encourage you to fill out the following form if you are interested in joining us: RSVP
The PEcAn Team will be at the 2016 American Geophysical Union (AGU) Conference and will be presenting the following talks and posters:
B12C-01: Mike Dietze "On the nature of prediction in ecology (Invited) ", 10:20 -10:35 Moscone West - 2004
B12C-02: Betsy Cowdery "Evaluating Productivity Predictions Under Elevated CO2 Conditions: Multi-Model Benchmarking Across FACE Experiments", 10:35 - 10:50 Moscone West - 2004
B12C-06: Istem Fer "Linking big models to big data: efficient ecosystem model calibration through Bayesian model emulation", 11:35 - 11:50 Moscone West - 2004
B24A-03: Michael Dietze "Show me the data: advances in multi-model benchmarking, assimilation, and forecasting", 16:30 - 16:45 Moscone West - 2004
B21B-0438: James Simkins "Partitioning sources of uncertainty in projecting the impact of future climate extremes on site to regional ecosystem carbon cycling", 08:00 - 12:20 Moscone South - Poster Hall
B31K-01: Alexey Shiklomanov "Synthesizing trait correlations and functional relationships across multiple scales: A Hierarchical Bayes approach", 08:00 - 08:15 Moscone West - 2010
B51I-05: Alexey Shiklomanov "Leaf optical properties shed light on foliar trait variability at individual to global scales (Invited)", 09:00 - 09:15 Moscone West - 2006
B52A-03: Toni Viskari " Which factors limit constraining ecosystem models with remote sensing observations? ", 10:50 - 11:05 Moscone West- 2006
Thanks to Abby Swann for having Mike Dietze out to UW to give a talk about PEcAn. It was great to catch up with Abby, Janneke HilleRisLambers, and Lauren Buckley, as well as to meet a diverse set of faculty and students.
Members of the PEcAn and PalEON modeling teams decendend on UC Berkeley for 3 intensive days of programming aimed at getting state data assimilation (SDA) working for paleoecological data. While some of these models use other data assimilation technology (e.g. CLM-4.5 is already coupled to DART), most attendees were focused on getting models into PEcAn's SDA scheme. Specifically, Alexey Shiklomanov and Tony Gardella worked on coupling ED2 and Istem Fer worked on coupling LPJ-GUESS. At the same time Christy Rollinson worked on generating an ensemble of meteorological inputs while Ann Raiho worked on adjusting the SDA code to handle that ensemble and pass a different met driver to each ensemble member. Mike Dietze helped with system design and started tackling the met workflow's handling of the met ensemble. Everyone found it a productive experience and went home with a lot of home work still left to do before everything's working.
We're also excited to report on discussions with Sean McMahon (Smithsonian) about developing the capacity to ingest data from the global CTFS/ForestGEO plot network. This network spans 63 large plots and over 6 million trees from the tropics to the boreal. ForestGEO represents an invaluable benchmark for ecosystem models, especially in the tropics, and we look forward to working with Sean and others on developing innovative multi-model / multi-site analyses. Be aware that not all ForestGEO site data is public, so only a subset of sites will be supported initially, but we'll be developing tools that will allow all sites in the network to easily load their data and leverage PEcAn in their research.
Excited to report on discussions with John Battles about his plans to stream the entirety of the Hubbard Brook data catalog into PEcAn. John is particularly excited about the ability to update models continuously as new data comes online every year. We will be working with John on ingest data this spring during his sabbatical. To meet this demand, we'll be doing a major overhaul of our data ingest documentation and tweaking some of our tools, all of which is aimed at benifiting the larger community (John's just volunteered to be our guinea pig).
Thanks to Josh Fisher for having Mike Dietze out to NASA JPL to give a talk about PEcAn and to really dig into the nitty-gritty details of how the JPL Model Farm works. We're excited to report that we see layed out a path forward for incorporating Model Farm technology into PEcAn in order to produce regional and global model runs. We're excited to be collaborating with Josh and his team and will keep everyone posted on progress toward making PEcAn global!
Mike Dietze gave a talk about PEcAn to the Earth System Science program at UC Irvine. Thanks to Jim Randerson for having us down and all the useful discussions with students and faculty. Exciting to see progress on ILAMB and talk about ways to work together.
Mike Dietze gave a talk at LBNL on PEcAn, PalEON, and our new forecasting initiative.
PEcAn lead PI, Mike Dietze, is spending the fall at Lawrence Berkeley National Lab working with the NGEE-Tropics team to get CLM-FATES (formerly known as CLM-ED) into PEcAn and get the NGEE-Tropics team set up for using PEcAn in their exciting work bringing together models and data in the tropics. Look for FATES support in the next PEcAn release!
Mike Dietze, Ann Raiho, and Christy Rollinson all taught at this year's PalEON summer course on assimilating paleo date into models. PEcAn post-doc Istem Fer was also there as a student. In this one week course students work through an exciting case study in how PEcAn accelerates our ability to assimilate data into models. Students started the week coring trees at the UNDERC field station, then prepped and read the cores. Within PEcAn students then did basic model runs with SIPNET for UNDERC and then sensitivity and uncertainty analyses to determine which parameters had the largest impact on NPP. Next, SIPNET was calibrated against the near-by Sylvania flux tower using PEcAn's parameter data assimilation tools. Finally, the tree ring data were assimilated into SIPNET using PEcAn's state data assimilation tools, accounting for a range of uncertainties estimated by the data.land and allometries modules. It's amazing to see students assimilate data they've just collected -- it's a great demonstration of the possibilities for near-real time model-data synthesis.
Camp PEON: Assimilation, multi-site, soil params, Maespa, LPJ-GUESS, improvements to web & tutorials
State Data Assimilation (SDA)
One of the main goals of this release was to re-generalize the EnKF and generalized ensemble filter that @araiho has been using with LINKAGES back to also working with SIPNET and to bring the design one step closer to working with many models. This is working (yeah!), but not yet fully generalized.
The revised SDA tutorial has been migrated to documentation/tutorials/StateAssimilation and the gitbook updated to explain the use of SDA tags in the settings file and the main functions used.
The basic PEcAn workflow was substantially re-organized to allow PEcAn to more easily run multiple sites at once (see Site Groups below). The core of the design is to split PEcAn’s settings object (what’s read from the XML) into a list of settings objects, one for each run, and applied to existing functions. Both Settings and SettingsList are now R classes. Within the settings file itself, there is now a
Note: Multisite/multimodel run SettingLists cannot yet be generated through the web interface.
Parameter data assimilation
Generalized to multiple PFTs
Emulator now supports a range of additional features that can be controlled through the XML settings:
Non-PFT parameter constraint
As an experiment in giving PEcAn control of non-PFT parameters, we've created a 'soil' PFT for SIPNET that contains soil biogeochemical parameters. If successful, we will extend this approach to other models. This change already improves model calibration and validation and has allowed us to assess the relative contribution of plant vs soil parameters to model output uncertainty.
Maespa has basic functionality (thanks @mdekauwe for all the help!) LPJ-GUESS running, but without PEcAn controlling parameters
Web InterfacePEcAn landing page updated to be links to tools. Avoids confusion when users forget to add the /pecan, /bety, /rstudio, etc. to the web address.
Site groups has been populated with a number of major research networks: Ameriflux, Fluxnet, Fluxnet 2015 Tier 1 (open), LTER, NEON, ForestGEO, PalEON
Links to documentation, gitter, and bug reports from all pages
Additional mouseover text and links to model and variables tables
Legend in default plot, removed spline from Shiny plot
Camp PalEONThis is a special release for the PalEON summer course, and comes pre-installed with extra libraries and data for the course
TutorialsIn addition to migrating the PDA and SDA tutorials to RMarkdown, the model-data comparison at the end of Demo 1 has also been migrated to documentation/tutorials/AnalyzeOutput
BenchmarkingWhile still in alpha development, @bcow contributed a large chunk of code bringing us much closer to automated benchmark execution and a first pass at a Shiny page for benchmark visualization.
BugfixesC_stack bug in run.sensitivity.analysis: fixed on new VM but older users should upgrade gridExtra R package Units in figures and reading of model output #792 -- downstream applications/code should check for changes.
Can skip meta-analysis in pecan settings. May require updating old pecan.xml, particularly for integration tests.
Interface: met date checking
Misc: #906 #916 #919 #923 #948 #950 #951
Known IssuesThis version of PEcAn is misconfigured and is using the wrong server for Brown Dog. The server that is used is the development server for Brown Dog and might result in no data being returned. During the Camp PEON the development server will be using the same code as the production server. If you want to use BrownDog after Camp PEON we highly encourage you to sign up with Brown Dog. We will at that point contact you and send you the instructions to modify your VM to use your new account and the production servers.
CMIP5, Shiny, FLUXNET2015, Global Sensitivity
We’ve add support for running models into the future under climate change scenarios! This release adds support for grabbing GFDL runs for all four RCP scenarios multiple experiments. RCP4.5 is the ‘default’, but Advanced Edit can be used to change the model (CM3, ESM2M, ESM2G), scenario (rcp26, rcp45, rcp60, rcp85), or ensemble member (r1i1p1, r3i1p1, r5i1p1). We are working with Earth System Grid to provide support for other climate models in future PEcAn releases.
We’ve added support for R Shiny visualizations in a way that we hope will allow users to easily add new Shiny apps to the catalog of PEcAn visualizations. Simply drop a new app into the pecan/shiny folder and it should go live. We’ve initially populated our Shiny catalog with two visualizations:
Similar to our current visualizations, which plots model output timeseries, but now shows output over all years and without a default 1-day time averaging.
Displays results from our new Monte Carlo global sensitivity analysis. Select outputs by model parameter (all model outputs), by output variable (all model parameters), or dig down into individual parameter-by-output relationships.
FLUXNET2015 and Ameriflux LBL:
At the site-level we’ve added support for the new FLUXNET2015 synthesis product, extending PEcAn’s support for flux tower sites globally. For Ameriflux, we’ve also added support for data on the new Lawrence Berkeley server. Past support for data on the ORNL server will also continue so long as that site remains active, but LBL is the preferred server and will contain the most up-do-date data.
Finally, in response to the above changes and feedback from users, we’ve tried to streamline the main PEcAn tutorial. The Parameter Data Assimilation (PDA) tutorial has also been updated to leverage emulator upgrades in the last release.
MAESPA, G’DAY, LPJ-GUESS not yet ready. JULES supported but not installed on VM, need to download and install yourself.
Met processing bugs: code for extending the date range from previous runs has a bug. Current workaround through advanced edit: set new met output directory (See PDA demo). For CRUNCEP and GFDL, workflow gets tripped up when extending to a new model. Currently requires running met process code by hand in R. Bugfixes will be in next release (Aug 12).
In the pipeline:
Next release will focus on PalEON summer course (SDA, tree ring & allometry modules, hopefully alpha support for PalEON met and NEOTOMA database). Also hope to bring Benchmarking up to a functional beta version, add GLDAS & NLDAS met.
Mike Dietze, Ankur Desai, and Betsy Cowdery are all returning to teach at this year's Flux Course. PEcAn grad students James Simkins (Wisc) and Tess McCabe (BU) are also in attendance. The Flux Course is one of the PEcAn team's big annual deadlines, so look for a bunch of new, cool features in release v1.4.7!
Mike Dietze presented an evening seminar on PEcAn at the Harvard Forest REU summer program, which included a hands-on demo. Dietze enjoyed getting to meet this year's crop of REUs over dinner and we wish them all the best of luck in finishing up their summer projects.
Mike Dietze's book "Ecological Forecasting" has been officially accepted for publication by Princeton University Press. Expect the book out Spring 2017. This book captures much of the philosophy and tools of the PEcAn project (check out the PEcAn case study in Chapter 12!), while also looking more broadly at the state of ecological forecasting in many ecological disciplines. This book grew out of the Broader Impacts of our NSF Advances in Biological Informatics (ABI) awards.
Mike Dietze attended the Ecological Society of America's SBI short course for project directors. Thanks to all the great instructors for providing with a lot to think about!
As a result of this experience, the PEcAn team has kicked off it's first ever strategic planning exercise. As a place to start, check out our new Vision and Mission Statements
Vision: Ecosystem science, policy, and management informed by the best available data and models
Mission: Develop and promote accessible tools for reproducible ecosystem modeling and forecasting
Mike Dietze again taught at PROFOUND spring school on Bayesian model calibration.
Additional information about the course can be found here
Ankur Desai, James Simkins, and Shawn Serbin represented the PEcAn team at the International Land Model Benchmarking (ILAMB) Workshop in Washington, D.C.
Serbin gave a talk entitled "PEcAn: A community tool to enable synthesis, evaluation & forecasting"
The PEcAn Workshop was held from May 9th-11th, 2016 at Boston University with the goals to:
This release has the beta release of MAESPA and LPJ-GUESS models, an upgrade to parameter calibration (emulator tools and bayesian sampler tools), and beta release of EnKF generalization. (You can find the VM here and release notes on Github)
Other changes include:
Mike Dietze presented on the PEcAn project at the North American Carbon Program (NACP) Workshop "Development of Predictive Carbon Cycle Science" in College Park, Maryland.
We've coupled another model and meteorological product to PEcAn. The PREdiction of Light use efficiency, Evapotranspiration, and Soil water (PRELES) model comes from our friends at the Natural Resources Institute Finland. The global CRUNCEP met product gives gives us the capability to run any model anywhere around the world! (You can find the VM here and release notes on Github)
Other changes include:
Shawn Serbin and Alexey Shiklomanov attended the Integrated Network for Terrestial Ecosystem Research on Feedbacks to the Atmosphere and ClimatE(INTERFACE) meeting in St.Pete's Beach, FL. Shawn led a working group on Remote sensing and both Alexey and Shawn presented posters.
Mike Dietze, Elizabeth Cowdery, and Toni Viskari represented our team at the Ecosystem Demography(ED) meeting in Boulder,CO. Mike gave a presentation on PEcAn and all three participated in various working group discussions. One discussion specifically surrounded the coupling of PEcAn and CLM-ED. It was a productive discussion and we will continue working on this endeavor with our friends at NCAR.
Mike Dietze and Andy Fox organized a meeting on "Operationalizing Ecological Forecasting" in Ft. Collins, CO during the first week of January. It was hosted by the USGS Powell Center and funded by NEON. Mike gave a talk titled "The PEcAn Project: Putting Ecosystem Model-Data Fusion in your Pocket". More information about the meeting can be found here.
The PEcAn Team is gathered at the 2015 American Geophysical Union Conference and will be presenting the following talks and posters:
B11C-0441: Betsy Cowdery "Assessing model sensitivity and uncertainty across multiple Free-Air CO2 Enrichment experiments", Monday, 14 December 2015 08:00 - 12:20 Moscone South- Poster Hall
B11C-0448: Ann Raiho "Assimilating Paleoecological Data into a Forest Gap Model", Monday, 14 December 2015 08:00 - 12:20 Moscone South- Poster Hall
B11L-02: Josh Mantooth "Tree Nonstructural Carbohydrate Reserves Across Eastern US Temperate Forests", Monday 14, December 12, 2015 8:15-8:30 Moscone West-2004
B11N-06: Ryan Kelly "Widespread Burning of Alaskan Boreal Forests Overcomes Fuel Limitation in the 21st Century", Monday, 14 December 2015 09:15 - 09:30 Moscone West- 2010
B11N-07: Mike Dietze "Fusing data and models to forecast disturbance impacts on ecosystems: past, present, and future", Monday, 14 December 2015 09:30 - 09:45 Moscone West- 2010
B14B-01: Shawn Serbin "Reducing the uncertainty in the projection of the terrestrial carbon cycle by fusing models with remote sensing data", Monday, 14 December 2015 16:00 - 16:15 Moscone West- 2004
IN21E-07: Mike Dietze "Chasing the long tail of environmental data: PEcAn is nuts about Brown Dog", Tuesday, 15 December 2015 09:30 - 09:45 Moscone West- 2020
B42A-04: Ankur Desai "Detection of Extreme Climate Event Impacts to Terrestrial Productivity From Airborne Hyperspectral Imagery", Thursday, 17 December 2015 11:05 - 11:20 Moscone West- 2006
IN43D-08: David LeBauer "Reusable Software and Open Data Incorporate Ecological Understanding To Optimize Agriculture and Improve Crops", Thursday, 17 December 2015 15:25 - 15:40 Moscone West- 2020
B53D-0594: Alexey Shiklomanov "Applications of spectral inversion to understanding vegetation functional trait relationships", Friday, 18 December 2015 13:40 - 18:00 Moscone South- Poster Hall
B53B-0549: Toni Viskari "Assessing the impact of radiative parameter uncertainty on plant growth simulation", Friday, 18 December 2015 13:40 - 18:00 Moscone South- Poster Hall
B53B-0544: Shawn Serbin "Linking vegetation structure, function and physiology through spectroscopic remote sensing", Friday, 18 December 2015 13:40 - 18:00 Moscone South- Poster Hall
Gearing up for a major release before the American Geophysical Union conference, we're bringing in a key new feature: the ability to launch workflows from the web interface and have them executed on remote machines ! (You can find the VM here and release notes on Github)
Other features include:
Mike Dietze presented an overview of PEcAn's trait modeling research as part of a Department of Energy workshop assessing how trait-based approaches can inform ecosystem and earth system models. Dietze's presentation, can be found here.
Mike Dietze visited the University of Georgia's Odum School of Ecology on October 5-7 and talked about ongoing work on the PEcAn project in a talk entitled Forecasting Ecosystems: Challenges and Opportunities
Mike Dietze presented the talk “Breaking the modeling communication gaps: models talking with ecologists, the data, and each other.” at the centennial Ecological Society of America Annual Meeting held August 2015 in Baltimore, MD. This talk was part of an excellent session "Creative Approaches for Addressing Ecological Uncertainty in Earth System Models" organized by Nick Smith and Jeff Dukes
The DOE Advanced Research Projects Agency-Energy (ARPA-E) Transportation Energy Resources from Renewable Agriculture (TERRA) program announced two awards to Illinois and NCSA that will support the further development of PEcAn and BETYdb to support the use and development of a new cyberinfrastructure to synthesize sensor data and develop automated platforms for identifying promising genotypes in breeding trials. Read more about the big data and computing 'reference' platform that will provide a core set of phenotyping data and the development of an all-terrain robot . These projects aim to efficiently quantify plant traits and predict yield potential of individual plants using sensors that can be deployed on robots, tractors, and drones.
We are looking for help writing an R based interface to the BETYdb API.A few functions have already been added to the rOpensci traits package that can be found in the file R/betydb.R. If you can help resolve open issues, implement new functionality, and write vignettes please contact firstname.lastname@example.org, submit a pull request or review our current short-term job announcement
Just in time for Flux Course PEcAn 1.4.3 is released. As always you can download the latest VM from and full release notes can be found on Github. The PEcAn VM has also been released on Amazon AWS as a public AMI.
Highlights of this release are:
We are excited to announce that Tony Gardella will be joining the PEcAn team as Project Manager starting August 1st. He is a recent graduate of Boston University with a Bachelor of Arts degree in Environmental Analysis and Policy. His passion for solving environmental issues and interest in the tools necessary to do so brought him to the PEcAn project. We are confident that his experience in the social sciences, understanding of the natural sciences, and computer science skills will be able to facilitate the growth of the project by engaging a wider user community and improving its accessibility. He cannot wait to work with you all and hopes to do so soon.
Mike Dietze, Ankur Desai, and Betsy Cowdery will be representing the PEcAn team at this year's edition of the annual Flux Course at Niwot Ridge, July 20-31.
The PEcAn team is furiously working on new features, tutorials, and R vignettes to show off at the flux course. Because of all this work keep an eye out for PEcAn 1.4.3, the release of which will be timed for the start of camp!
Mike Dietze is presenting the following talk on PEcAn at the International Association for Landscape Ecology (IALE 2015) World Congress in Portland Oregon.
Mike Dietze once again presented the PEcAn project to the Harvard Forest REU (Research Experience for Undergraduates) summer students. In addition to the seminar and a bit of live demo, Mike enjoyed dinner and a wide ranging conversation with this year's crop of students.
Mike Dietze attended the Building Global Ecological Understanding Workshop at the University of Delaware, June 3-5, 2015. This workshop focused a lot on Big Data in ecology, with considerable discussion on how cyberinfrastructure, such as PEcAn and Brown Dog, can help address cutting edge questions in ecology.
Mike Dietze served as one of the instructors at a week-long spring school Bayesian calibration, forecasting and multi-model predictions of process-based vegetation models organized as part of the COST PROFOUND project. This course ran 12.-16. of May and was located in Rencurel, France, in the foothills of the French Alps near Grenoble. In addition to presenting a hands-on tutorial in PEcAn, Mike also gave lectures on Maximum Likelihood and Bayesian Prediction/Forecasting.
Mike gave a talk in the Biological Sciences Department at the University of New Hampshire on May 1st entitled "The PEcAn Project: A Community Platform for Synthesis & Forecasting of Ecosystems". Mike's host, Kirk Broders, was unfortunately out of town, but Scott Ollinger showed Mike around campus. Mike, Scott, and Scott's lab also enjoyed a long conversation about PEcAn, PNET, and the ongoing forest ecology research across New England.
Mike Dietze gave a seminar in the Ecology and Evolution Department at Stony Brook University on April 29th entitled "The PEcAn Project: A Community Platform for Synthesis & Forecasting of Ecosystems". Thanks to Mike's former Moorcroft Lab labmate, Heather Lynch, for serving as his host and thanks to everyone as SBU for two days of fun conversations and an awesome departmental t-shirt
By sheer coincidence the email from the BU Dean awarding Mike tenure arrived DURING his seminar. This is definately one campus visit that he'll remember for a long time!
SummaryThe Ecological Forecasting lab at Boston University seeks a full-time software project manager to coordinate phase 2 development, user engagement, outreach, and education on the PEcAn project (Predictive Ecosystem Analyzer, www.pecanproject.org). PEcAn is an ecoinformatics system designed to make ecosystem models more accessible to the research community and managers, to manage the flows of information into and out of such models, and to facilitate analysis, data assimilation, visualization, and forecasting. The project manager would work closely with Prof. Michael Dietze and project collaborators at the University of Wisconsin, the National Center for Supercomputing Applications, and Brookhaven National Lab.
Required Education, Training, and Skills
Compensation$35k/year salaried Visiting Fellow position with benefits and vacation Annual contract renewable up to 4 years.
To apply send cover letter, CV, and contact info for 3 references to email@example.com
Shawn Serbin, Toni Viskari, and Alexey Shiklomanov will be presenting posters at the NASA workshop on the work being done in PEcAn to couple ecosystem models to remotely-sensed data. In particular, check out Alexey's awesome work in developing the PEcAn RTM (Radiative Transfer Model) module. This modules is the basis for Alexey's hierarchical Bayes inversion of the PROSPECT leaf-level spectra model and linages between leaf economic and spectral traits.
The Brown Dog team is making a big presentation at NSF Headquarters about all the great tools being developed to make data more accessible and make data conversions and extractions easier. Mike Dietze is presenting work done by Betsy Cowdery, Rob Kooper, and others to develop the Brown Dog powered meteorology workflow in PEcAn. Look for more details in the next PEcAn version release (coming soon!). Also presenting some early work by Josh Mantooth on the start of the PEcAn vegetation data standards.
Mike Dietze presented PEcAn in a departmental seminar at Montana State. Thanks to Paul Stoy for hosting Mike in his visit to Montana State and to all the folks that came out to learn about PEcAn. Particularly exciting was a long discussion with Ben Poulter's lab about ways that PEcAn might contribute to their work with LPJ, LPJ-Guess, and ORCHIDEE
Adrien Finzi and Mike Dietze presented work at the Harvard Forest Annual Meeting on work to construct composite allometric equations for the dominant tree species at Harvard Forest using the PEcAn allometry module. This module will sythesize over multiple allometric equations and raw data to generate Hierarchical Bayes estimates of parameters, uncertainties, and site-to-site variability. This module can be applied to any individual species, species grouping, or plant functional type. Look for more from an in prep manuscript from Finzi, Dietze, and graduate student Josh Mantooth
The PEcAn team will have a large presence at the North American Carbon Program's 5th All Investigator's Meeting in Washington DC, January 26-29, 2014.
On Tuesday, 3:12 pm Mike Dietze will be presenting a talk "The PEcAn Project: a scalable, multi-model platform for uncertainty quantification, analysis, and propagation" in the Uncertainty session
We will be organizing a hands-on PEcAn workshop at 7pm on Tuesday. Details here
Also look for poster presentations by Betsy Cowdery, Josh Mantooth, Afshin Pourmokhtarian, Christine Rollinson, and Alexey Shiklomanov
The PEcAn team will be presenting the following talks and posters at the 2014 Fall meeting of the American Geophysical Union in San Francisco:
B13C-0202 "The PEcAn Project: Accessible Tools for On-demand Ecosystem Modeling" Monday, December 15, 201401:40 PM - 06:00 PM Moscone West Poster Hall
B13C-0192 "Constraining carbon budgets at a regional scale: fusing forest inventory data with a cohort-based biosphere model" Monday, December 15, 2014 01:40 PM - 06:00 PM Moscone West Poster Hall
B13C-0198 "A Hierarchical Analysis of Tree Growth and Environmental Drivers Across Eastern US Temperate Forests" Monday, December 15, 201401:40 PM - 06:00 PM Moscone West Poster Hall
B52A-04 "Caught in the flux net: disentangling error, uncertainty, heterogeneity, and spatial process in biogeochemical scaling (Invited)" Friday, December 19, 2014 11:05 AM - 11:20 AM Moscone West 2002
B51G-0101 "Scaling Forest Management Practices in Earth System Models: Case Study of Southeast and Pacific Northwest Forests" Friday, December 19, 201408:00 AM - 12:20 PM Moscone West Poster Hall
B53D-0218 "Measured and modelled carbon and water fluxes in hybrid willows grown for biofuel production" Friday, December 19, 2014 01:40 PM - 06:00 PM Moscone West Poster Hall
B53I-07 "Integrating satellite and tower phenology: a case-study in real-time ecological forecasting" Friday, December 19, 201403:10 PM - 03:25 PM Moscone West 2002
OverviewThis version supports conversion of met files. The code can now download and convert the met data to CF standard and to model specific format. The met data will along the way be gap filled. This allows any researchers to pick any Ameriflux site and run the SIPNET model on this site. As always the latest VM can be downloaded from NCSA
pecan.xml ChangesThe met tag can now support a special notation that lets it download and convert the met data. The met tag can have an input="Ameriflux" and output="sipnet" attribute. Currently these are the only supported conversion but more will be added in the next release of PEcAn.
web changesWhen running in the web interface you now have the ability to edit the pecan.xml before the software runs. This allows you to change the number of ensemble runs done, as well as any other changes to the pecan.xml. After the run finishes, you can now download the sipnet clim file used for a particular run. You can now also plot different variables against each other instead of time vs one of hte variables.
Mike Dietze gave the keynote address, "Models and Data", at the inaugural meeting of PROFOUND (FP 1304 Towards robust projections of European forests under climate change), a new EU COST Action. This presentation on challenges in bringing models and data together focused heavily on PEcAn.
Mike will be particpating in a number of COST PROFOUND working groups and task groups. In particular, Mike Dietze and David Cameron (UK) will be co-orginizing Task Group 15 - Statistical problems in fitting models to heterogeneous data
Mike Dietze gave a seminar "The PEcAn Project:(A Community Platform for Synthesis & Forecasting of Ecosystems" in the Department of Ecology, Evolution & Environmental Biology (E3B) at Columbia University. Thanks to Maria Uriarte for hosting!
Mike Dietze gave a seminar in the Biology Department at Kent State titled "Terrestrial Ecosystems: Past, Present, & Future"
While at Kent, Mike also ran a two-hour PEcAn workshop for departmental grad students and faculty.
Thanks to Chris Blackwood for being a great host for both the seminar and workshop!
Mike Dietze presented his tenure seminar errestrial Ecosystems: Past, Present, & Future" to the Department of Earth and Environment at Boston University. This talk focused on the PalEON and PEcAn projects and layed out the vision for PEcAn's future direction.
This version supports new model types better. BETY now allows to add new model types to the database. You can associate these model types with PFT's. Each model type will have a list of required and optional inputs associated. These inputs will be listed in <run><inputs> tag.
As always the latest VM can be downloaded from http://isda.ncsa.illinois.edu/download/index.php?project=PEcAn&sort=version
Support for DALEC
A new model has been added to this release, DALEC. BETY already has a modeltype DALEC in the database to support this model.
Using the new modeltypes we can now check the inputs specified in the pecan.xml file. All inputs should now be listed under the <run><inputs>. The code will try and update your pecan.xml file to this new format. Any missing inputs that are required will quit the run, allowing you to quickly fix your pecan.xml file. For updated information see https://github.com/PecanProject/pecan/wiki/PEcAn-Configuration.
The database requires updates for the model types table. You will need to update BETY to be able to see this. This requires version 3.2 of BETYdb (https://github.com/PecanProject/bety/releases/tag/betydb_3.2).
Sites now store their information as geometry objects. If you have custom queries that use the site.lat and site.lon information you will need to change this to ST_X(ST_CENTROID(geometry)) AS lon, ST_Y(ST_CENTROID(geometry)) AS lat.
One major change is to the function called get.trait.data(). This function now requires a model type as the second parameter. The model type is used to make sure we get the right PFT. This allows multiple PFTs to have the same name, but be distinguishable by the combination of name and model type.
Mike Dietze, Ankur Desai, and Kenton McHenry submitted the PEcAn renewal proposal "Collaborative Research: ABI Development: The PEcAn Project: A Community Platform for Ecological Forecasting" to the NSF program Advances in Biological Informatics (ABI). The PEcAn 2 proposal focuses on three additional challenges:
David LeBauer presented a talk entitled "Combining heterogeneous data and process understanding using the Predictive Ecosystem Analyzer" on Wednesday, August 13, 2014 in Symposium 11: Challenges and Advances in Statistical Software For Ecology. Shawn Serbin and Mike Dietze were co-authors on the presentation.
Dietze also presented a talk entitled "Predicting phenology: A case-study in real-time ecological forecasting" on Monday, August 11, 2014 in Organized Oral Session 4: The National Ecological Observatory (NEON): Opportunities and models for building synergistic partnerships with the community to advance continental scale ecology.
This version focuses on more database updates.
As always you can download the VM from: http://isda.ncsa.illinois.edu/download/index.php?project=PEcAn&sort=category
the script to add sites is now part of PEcAn. To add the sites data you can use pecan/scripts/add.data.sh the script to add models is updated and will setup the binaries and is called pecan/scripts/add.models.sh Both scripts can be called multiple times, but will only add the information to the database once.
In pecan.xml the database section now has 2 subsections, bety and fia. Any existing information that was in the database section in pecan.xml will need to be moved to the bety subsection, read.settings() will automatically update your pecan.xml to reflect this. The <bety><write> option in the pecan.xml is now part of the <database><bety> section, read.settings() will automatically update your pecan.xml to reflect this. The ED land use/soil/vegetation/thermal sums data files are now inputs in BETY that need to be specified as inputs.
Action Required: add ED data files to database, see GitHub PEcAn Issue #120
Support for meteorology conversion from netCDF CF standard to SipNET's .clim driver files PEcAn Web App
config.php now declares the fully qualified host name ($fqdn)
the config.php in pecan/web will have new database configuration parameters, please see config.example.php. no longer needed to specify vet/soil/ed_inputs in config.php, these are now stored as inputs in BETY. if the fia database is specified in config.php, the user can now select fia2ed from psscss selector. Action Required: update pecan/web/config.php, see pecan/web/config_example for latest settings.
To update BETY database either download a new version or update the database.
In the database there is now a link from a model to the dbfiles table. This allows you to have a single model, with binaries located in different places on different systems. Each PFT now stores the model type associated with it, the web application will use this to find the right pfts to display for the user, and read.settings will check to see if the PFT is compatible with the model.
Action Required: add binaries to models, for example using the pecan/scripts/addmodels.sh.
Action Required: the migration of the database will add the appropriate model_type to each PFT based on the prefix of the PFT.
Kenton McHenry (Brown Dog PI, PEcAn co-PI) hosted the Brown Dog Early User Workshop from July 22-13, 2014 at the National Center for Supercomputing Applications in Urbana, Illinois. Brown Dog is a new, NSF-sponsored cyberinfrastructure initiative to make unstructured, uncurated data more accessible to scientific community and the general public. Brown Dog consists of two key tools, the Data Access Proxy, which converts file formats, and the Data Tilling Service, which extracts useful information and meta-data.
Mike Dietze (PEcAn PI, Brown Dog co-PI) gave a presentation (pdf, YouTube) on how Brown Dog will be used to access and harmonize vegetation and meteorological data, and how it will be incorporated into PEcAn. Moving forward, Brown Dog will play a key roll in the PEcAn project, helping to manage flows of information into PEcAn and proccessing data into model drivers.
Rob Kooper (NCSA Senior Programmer on both Brown Dog and PEcAn) presented on how Brown Dog will interact with other NSF computational resources. David LeBauer (PEcAn) presented a lightning talk on how Brown Dog might benifit BETY-db, the database that underlies PEcAn.
Mike Dietze was invited back to the Harvard Forest Summer REU program again this year to present an evening seminar "The PEcAn Project: Carbon-Cycle Reanalysis Facilitated by Model-Data Ecoinformatics"
The following paper has been published
Niu S, Y Luo, M Dietze, T Keenan, Z Shi, J Li, FS Chapin III. 2014. The role of data assimilation in predictive ecology. Ecosphere 5:art65. http://dx.doi.org/10.1890/ES13-00273.1
Switch to PostgreSQL, enable distributed instances of BETYdb, maintaining local database changes
This version officially switches to postgreSQL. This allows for the script to enable to keep local changes to the PEcAn database.
This enables multiple instances to run BETYdb, and for local changes to be exchanged across databases. Each installation has a reserved set of foreign key ids (e.g. Boston University has 1,000,000,000-1,999,999,999 reserved. Edits to existing records added on another server will not be preserved.
Mike Dietze participated in the New Phytologist workshop "Improving Representation of Photosynthesis in Earth System Models" in Montauk, NY organized by Alistair Rogers, Belinda Medlyn, and Jeff Dukes. Mike presented on how PEcAn can help constrain photosynthesic traits in Earth System Models and assess the uncertainty in model projections. A meeting report was published in New Phytologist and can be found here
Xiaohui "Sunny" Feng successfully defended her PhD dissertation "PRODUCTIVITY, PHYSIOLOGY, COMMUNITY DYNAMICS, AND ECOLOGICAL IMPACTS OF A GRASSLAND AGRO-ECOSYSTEM" at the University of Illinois, Urbana-Champaign.
FORECAST RCN workshop, “Advancing Software for Ecological Forecasting”, March 25-27 2014, Urbana, IL
Mike Dietze was also in attendance
Dietze M, S Serbin, C Davidson, A Desai, X Feng, R Kelly, R Kooper, D LeBauer, J Mantooth, K McHenry, D Wang. 2014. A quantitative assessment of a terrestrial biosphere model's data needs across North American biomes. Journal of Geophysical Research Biogeosciences DOI: 10.1002/2013JG002392 link
Dietze M. 2013. Gaps in knowledge and data driving uncertainty in models of photosynthesis. Photosynthesis Research 19:3-14 DOI: 10.1007/s11120-013-9836-z
Mike Dietze presented the seminar talk "The PEcAn Project: Accessible ecoinformatic tools for carbon-cycle model-data analysis and assimilation" at the Ecosystems Center at the Marine Biological Laboratory in Woods Hole, MA.
The proposal “Assimilation of imaging spectroscopy data to improve the representation of vegetation dynamics in ecosystem models” lead by Shawn Serbin, with Mike Dietze and Phil Townsend as co-PIs, has been recommended for funding by NASA. This project will focus on assimilating hyperspectral data in the PEcAn framework by building a Radiative Transfer Model into the Ecosystem Demography model
The following PEcAn presentations occurred at the American Geophysical Union Fall Meeting, December 2013, San Francisco
Viskari, Dietze, Desai "Model-data assimilation of multiple phenological observations to constrain and forecast leaf area"
Serbin, Dietze, Desai, LeBauer, Viskari, Kooper, McHenry, Townsend "Assimilation of Leaf and Canopy Spectroscopic Data to Improve the Representation of Vegetation Dynamics in Terrestrial Ecosystem Models"
The following paper describes the statistical model underlying the PEcAn photosynthesis module
Feng X, M Dietze. 2013. Scale-dependence in the effects of leaf economic traits on photosynthesis: Bayesian parameterization of photosynthesis models. New Phytologist 200(4): 1132–1144 DOI: 10.1111/nph.12454 link
GE585 – Ecological Forecasting and Informatics was taught by Mike Dietze at Boston University in Fall of 2013
The course materials are available on GitHub
LeBauer and Dietze hosted a lunch-time workshop
"Combining Field Measurements and Ecosystem Models"
In addition, the following PEcAn presentations occurred at the Ecological Society of America Annual Meeting, Minneapolis, MN
Dietze “Assimilating forest inventory data into models”
LeBauer, Dietze, Jaiswal, Kooper, Long, Serbin, Wang "Reducing uncertainty through data-driven model development"
Hardiman, Serbin, Mantooth, Kennedy, Dietze "Assimilating PALSAR remote sensing data to reduce uncertainty in ED2 predictions of forest biomass dynamics following disturbance"
Kooper R, K McHenry, M Dietze, D LeBauer, S Serbin, A. Desai. 2013. Ecological Cyberinfrastructure and HPC Towards More Accurately Predicting Future Levels of Greenhouse Gases. Proceedings of XSEDE13: Extreme Science and Engineering Discovery Environment
LeBauer D, M. Dietze, B. Bolker. 2013. Translating Probability Density Functions: From R to BUGS and Back Again. R Journal Vol. 5/1, June, 207-209
LeBauer et al. 2013 was recommended in two different reviews on F1000. You can read the reviews at http://f1000.com/prime/718024607
Dietze was invited back to Harvard Forest to present PEcAn to the summer REU students (June 12)
Some of our favorite quotes:
“I learned how data and models can be used to predict the ecological impacts of climate change, and
￼how data can be fed back into models in an iterative process that increases our understanding of the
￼mechanisms behind observed patterns and enables us to make better predictions. I learned about
￼Bayesian statistical analysis and how parameter distributions are constrained by real-world data to
￼“I learned about ways in which my inquiries as a scientist could be better guided by methods I was
￼previously uninterested in.”
￼“That models should be made more accessible and usable to people other than the engineers of the
￼models. (I’m not going to lie; some of it went over my head a bit, however he was a very interesting
“I had no idea before this workshop how useful modeling is to scientific progress.”
Mike Dietze presented talks on the PEcAn project at the Boston University EBE Seminar series (April 1), Harvard University's ClimateTEA (April 2), and at Harvard Forest (May 2).
Seeking web programmers to develop PEcAn (pecanproject.org), BETYdb (betydb.org) front-end and back-end features. These tools are used to manage information and generate predictions of ecosystem functioning.
Requirements: Strong skills in one or more of: PHP, SQL, Ruby, R, or C, Fortran.
All applicants should review the PEcAn wiki at github.com/pecanproject/pecan/wiki. It is a large project, so don't get overwhelmed with the underlying science or the system design. But take the time to find a piece of code in your preferred programming language that you understand and could contribute to. Also look at 'foreign' language code - R, Ruby, C, Fortran that you would be interested in learning.
Please contact someone on the People page for more information.
Energy Biosciences Institute
Institute for Genomic Biology
University of Illinois at Urbana-Champaign
We seek a scientific web application programmer to join the ecosystem modeling program in the Energy Biosciences Institute at the University of Illinois in Urbana Champaign. This is a 12-month non-tenure track Academic Professional position with possibility of renewal. Our group uses data syntheses and ecosystem simulation modeling to provide ecological forecasts to scientists as well as non-specialists. An important objective of our work is the development of user-friendly web interfaces to the computational tools that we are developing.
RESPONSIBILITIES: The programmer will work with our team to develop web resources for ecological data management, modeling, and statistical inference. The programmer will assume responsibility for the design, implementation, and maintenance of front-end and back-end software that will provide into effective interfaces and data visualizations for two software tools that we are developing.
The programmer will be expected to provide leadership in the development of new software and the adoption of best-practices to support development and quality control, to remain current on the latest web technologies, to apply new knowledge to a cutting edge system, and to provide leadership in web-based front-end design and implementation decisions.
• Programming experience with web interface/back end design using web framework systems.
• Experience with or willingness to learn Ruby, PHP, R, SQL, Ruby-on-Rails, R, GIS, MVC and REST
To Apply: Please create your candidate profile athttp://jobs.illinois.edu and upload your letter of interest (including email address), resume and contact information for three professional letters of reference by March 18, 2013.
The PEcAn team will play a number of roles at the upcoming 4th North American Carbon Program All-Investigators Meeting, 4 – 7 February 2013 in Albuquerque, New Mexico.
Most directly relevant to the PEcAn mission, Mike Dietze, Ankur Desai, Trevor Keenan, and Quinn Thomas will be chairing a breakout session "Harnessing the 'long tail' of ecosystem carbon cycle observations: Approaches and challenges in synthesizing and assimilating non-automated and experimental data". This session is intended to serve as a call-to-arms for a “community” approach to model calibration, data assimilation, evaluation, and improvement. Tools such as PEcAn play a critical role in achieving this goal but will only be effective if this new approach to the communication between models and data (and between modelers and empiricists) is embraced by the scientific community.
In addition, Desai and Dietze will serve as chairs for two of the six presentation sessions, "Theme 1: Diagnosis of the Atmospheric Carbon Cycle" and "Theme 4: Ecosystem Impacts of Change" respectively.
Dr. Toni Viskari will be joining the PEcAn team as a postdoctoral fellow in the Dietze lab starting January 1, 2013. Toni will be working on our NSF ABI project using the northern Wisconsin "ChEAS" project as a testbed for regional-scale multi-proxy data assimilation. Toni joins us from the and the Finnish Meteorological Institute and the University of Helsinki in Finland, where he recently defended his dissertation "Estimating particle number size distributions from multi-instrument observations with Kalman Filtering".
The following PEcAn related talks and posters were presented at the 2012 meeting of the American Geophysical Union, December 3-7.
B22B-08. The modeled effects of fire on carbon balance and vegetation abundance in Alaskan tundra
Michael C. Dietze; Carl D. Davidson; Ryan Kelly; Philip E. Higuera; Fengsheng Hu
B34B-04. What do we need to measure, how much, and where? A quantitative assessment of terrestrial data needs across North American biomes through data-model fusion and sampling optimization (Invited)
Michael C. Dietze; Carl D. Davidson; Ankur R. Desai; Xiaohui Feng; Ryan Kelly; Rob Kooper; David S. LeBauer; Joshua Mantooth; Kenton McHenry; Shawn P. Serbin; Dan Wang
B41C-0296. A Web Interface for Eco System Modeling
Kenton McHenry; Rob Kooper; Shawn P. Serbin; David S. LeBauer; Ankur R. Desai; Michael C. Dietze
B41C-0306. The use of informed priors in ecosystem model-data fusion
David S. LeBauer; Michael C. Dietze
Dietze, M.C., D. LeBauer, R. Kooper. On improving the communication between models and data. Plant, Cell, and Environment in press DOI: 10.1111/pce.12043 abstract & pdf
This is a review and perspective piece for a special issue on plant models in PC&E. While it is a review, it highlights PEcAn as an important case study and summarizes much of the motivation for what we are trying to do. The paper is also a call-to-arms for a community-based approach to model-data synthesis that envisions a larger role for experimentalists in assimilating data into models, using models to express non-trivial quantitative hypotheses, and identifying and reducing the dominant sources of uncertainty in their systems.
Wang, Dan, David LeBauer, and Michael C. Dietze. In press. Predicting yields of short-rotation hybrid poplar (Populus spp.) for the contiguous US through model-data synthesis. Ecological Applications. http://dx.doi.org/10.1890/12-0854.1 abstract
In this paper the PEcAn workflow was used to calibrate the Ecosystem Demography model (ED2) for hybrid poplar using a combination of trait data and yield data. This analysis involved a trait meta-analysis, the use of sensitivity analysis and variance decomposition to identify dominant parameters, and the assimilation of annual yield data for three planting densities to further constrain these parameters. ED2 was then validated at eight additional sites distributed across the continental US, each with a range of genotypes and planting densities. Finally, the validated model was used to make a 0.5 degree projection of potential poplar yields and optimal rotation periods across the continental US and to make specific site-level projections about the changes in soil carbon relative to native forests and conventional agriculture. The most promising regions for poplar production are the Northeast, upper Midwest, and northern California, where projected poplar yields equal or exceed yields projected for Miscanthus, a highly productive perennial C4 grass being used for biofuel production.
As of September 2012, PEcAn version 1.2.6 is the latest stable version. This version was released for the August 2012 summer course "Assimilating long-term data into ecosystem models" that taught methods advanced by the PalEON project. PEcAn 1.2.6 is being used for ongoing research and applications, while development of PEcAn 1.3 continues. It contains all of the modules described in "version 1.1", with added functionality, improved stability, and improved useability that represents an increasing community of scientists and developers:
The public website is https://ebi-forecast.igb.illinois.edu/redmine/projects/pecan-1-0
The PEcAn version 1.1 source code includes an R package with functions and R scripts used to run the PEcAn workflow. The PEcAn 1.1 Virtual Machine can reproduce all of the computational steps used in the manuscriptFacilitating feedbacks between field measurements and ecosystem models (LeBauer et al in press). These files are provided here as an archived reference and accessibility. For further research, we recommend using the latest stable version available at www.pecanproject.org.
[[note: this position has been filled]]
Model-data synthesis and forecasting across the upper Midwest: Partitioning uncertainty and environmental heterogeneity in ecosystem carbon
A post-doctoral position in ecosystem model-data synthesis and ecological forecasting is available in the Dietze lab in the Department of Earth and Environment at Boston University.
This project focuses on integrating eddy-covariance, remote-sensing, forest inventory data, and other ground-based measurements within the Ecosystem Demography 2.2 model to understand and forecast regional-scale ecosystem dynamics across northern Wisconsin. The project is part of the larger PEcAn project (http://pecanproject.org), which aims to make ecosystem models, data assimilation, and forecasting more accessible, automated, and repeatable.
As a test bed for the development and application of the PEcAn informatics tools, this project is focused on the temperate/boreal transition zone in northern Wisconsin, a region that is expected to show large climate change responses and is one of the most data-rich regions in the country. The tools developed here will enable us to partition carbon flux and pool variability in space and time and to attribute the regional-scale responses to specific biotic and abiotic drivers. The data-assimilation framework will partition different sources of uncertainty, which will enable a better understanding of which are limiting our inference, and provide a more complete propagation of uncertainty into model forecasts.
Minimum qualifications are a doctoral degree in plant ecology or a related ecological or environmental science. Experience with, or interest in learning, Bayesian statistics, ecosystem modeling, remote-sensing, and ecoinformatics tools would be valuable. Salary is commensurate with experience and qualifications. Up to two years and four months of funding is available.
Submit a cover letter, CV, and contact info for 3 references to Dr. Michael Dietze (dietze at bu.edu) For more information visit http://pecanproject.org and MailScanner has detected a possible fraud attempt from "www.life.uiuc.edu" claiming to be http://people.bu.edu/dietze
Mike Dietze will be presenting an invited talk on the PEcAn project at the FORECAST RCN meeting in Woods Hole, MA, October 9-11, 2012.
Shawn Serbin will be presenting a talk “From multiple datasets to predictive ecology: The PEcAn workflow and model-data assimilation” at the ForestSAT 2012 meeting in Corvallis, OR, September 11-14th in Corvallis, OR
The first manuscript describing the initial components of the PEcAn project has been accepted for publication
LeBauer, D.S., D. Wang, K.T. Richter, C.C. Davidson, M.C. Dietze. Feedbacks between measurements and ecosystem models. Ecological Monographs in press
This link will be updated once the preprint manuscript is posted.
Now that this manuscript has been accepted we anticipate releasing v1.1 of the PEcAn software and Virtual Machine very soon, so please stay tuned!
Mike Dietze, Shawn Serbin, and Ankur Desai each presented a number of short talks on the PEcAn project at the annual investigator's meeting
for the Chequamegon Ecosystem Atmosphere Study (ChEAS) held at the Kemp Biological Station, WI on June 27-29th 2012.
Agenda and slides can be found at: http://flux.aos.wisc.edu/twiki/bin/view/Main/ChEASMeeting2012
Carl Davidson successfully defended his Master's thesis on June 11th 2012 on "The modeled effects of fire on carbon balance and vegetation abundance in Alaskan tundra".
Carl's thesis represents an incredible case-study in the application of the PEcAn system. Carl used the PEcAn meta-analysis, sensitivity analysis, and variance decomposition in an iterative fashion to initially develop three tundra plant functional types (PFTs: graminoid, deciduous shrub, evergreen shrub) for the ED2 model for the tundra based on a literature synthesis. Based on this analysis Carl designed and executed a field campaign in 2011 at the Toolik LTER, a rare example of using model analysis to directly inform field research. Carl then again used PEcAn to assimilate his own field data as well as flux data from two Ameriflux sites. Finally, Carl applied the calibrated ED2 model to run model experiments aimed at disentangling the ecosystem feedbacks involved in post-fire recovery in the tundra.
The thesis is archived at the University of Illinois
PEcAn was be used for all aspects of the modeling and data assimilation activities that are part of the "Assimilating Long-Term Data into Ecosystem Models" week-long summer course being offered at the University of Notre Dame Environmental Research Center (UNDERC), Land O' Lakes, WI.
Mike Dietze presented a total of 12 hours of lecture, tutorial, and hands-on demos using PEcAn, covering aspects of workflows for model calibration and analysis, parameter data assimilation at the Sylvania flux tower, and state-variable data assimilation of tree ring data.
PEcAn allowed students to go from coring trees Monday, mounting them Tuesday, and analyzing them Wednesday, to calibrating SIPNET Thursday using trait databases and flux tower data and assimilating the tree ring data on Friday.