Meeting Notes 2021 Science - ESCOMP/CTSM GitHub Wiki

2021 CTSM science meeting notes

Dec 16

Dec 2

Updates and Announcements

  • ClimGrass project participants
    • single point simulations of a managed European C3 grasslands.
    • CO2 x warming x drought experiment w/ N fertilizations with 2-3 harvests annually.
  • Recent Tags dev063,
    • Makes NUOPC the default driver
    • Do we need office hours to get people up to speed?
  • NCAR-NEON tools released
  • LMWG meeting Jan 31-Feb 4, 2022. Remote.
  • Dec 16 meeting?

Xueli Hao

  • Data assimilation of leaf area index and biomass into CLM in the Arctic Boreal region

Nov 18

Updates and Announcements

  • recent tags
    • Nothing to announce, but ctsm5.1.dev062 will bring in nuopc as default
  • Wins:
    • 200+ forks of CTSM! STAR the repo to make Bill's day.
    • Danica and Negin drove a live demo of NCAR-NEON results with 100+ participants in recent workshop.
    • NCAR-NEON tools will be released Friday.
  • LMWG meeting Jan 31-Feb 4, 2022. Remote.
    • LMWG solicited talks? Suggested people to encourage participation.

LULCC tech note, Peter Lawrence

Nov 4

CTSM and CESM3 development timeline

  • Suggested timeline for CTSM development

  • Are we closer to being able to identify what CTSM6 will likely look like?

    • LULCC on hillslopes?
    • Global FATES parameterization (of some flavor)?

  • When do we make our first CTSM5.2 tag (and 'finalize' CTSM5.1)?

    • We have made progress on a number of accomplishments including FATES-SP, nuopc, +NEON and PPE branches.
    • Should we release hillslope hydrology with CTSM5.1?
    • Can we identify priorities and dates.
    • How do we document this 'minor' release?

  • How do we consider supporting and curating tools for model analyses?

    • Negin has great examples from NEON, Daniel has more for the PPE, ctsm-py also has some useful tools, and FATES likely does too?
    • These wouldn't be on the escomp/ctsm repo, but elsewhere on github.
    • How do we organize & navigate this?

Oct 21

Introduction and welcome

  • Katya Jay, CU Boulder & Niwot Ridge LTER

NUOPC features and changes for CESM and CTSM.

Oct 7

Brian Dobbins shares on CESM-Lab and running CTSM in a container.

  • ability to run regional or generic single point without downloading whole inputdata would be nice.
  • Brain and Negin suggested that maybe we can do this interactively using subset_data, putting datm_files on an ftp server for download.

Sept 23

CTSM Wins!

Updates on projects, accomplishments, and newly funded projects

Andy Wood presents on Water Forecasting

Sept 9

Update on FATES-Spitfire results from Jackie Shuman

Aug 31 – CTSM SSG meeting

Update from Dave Lawrence

Status update

Part of the CTSM effort is about bringing together NoahMP and CLM; this is a priority internally. However, as Jean-Francois points out, it's important (especially externally) to discuss this in more broad terms.

There have been recent developments in adding flexibility to the surface characterization. Among other things, this includes some recent work by Sean Swenson to run CTSM with grid cells as catchments; adding dynamic lake area; runtime flexibility in the complexity of surface heterogeneity and number of soil and snow layers.

Range of complexities, including:

  • CTSM5(NWP)
  • CTSM5(CLMSP)
  • CTSM5(CLMBGC, BGC-crop)
  • CTSM5(FATES) – experimental
  • CTSM5(HP) – hydrologic prediction; TBD

Climate & weather applications share same physics. Can choose between versions 4.5, 5.0 or 5.1 physics.

Noah-MP has also had some recent developments (potential migration to CTSM):

  • Snow process & parameter updates to reduce biases in SWE & winter temperature
  • Incorporate vegetation-dependent wind-canopy absorption parameter based on Goudriaan (1977)
  • New parameterization of agriculture tile drainage and three irrigation methods: sprinkler, micro, and surface flooding with new irrigation input maps for CONUS
  • Crop model parameter updates
  • Various urban model updates

So we still have the situation where CTSM and Noah-MP are being developed separately; we still have to make some progress to get to the point where we're all collectively benefiting from our model developments.

Some big CTSM developments:

  • A lot of emphasis is being put into the transition to FATES as our vegetation model option for CESM3
    • Jean-Francois asks if this is being done in a way that one could fairly easily bring in an alternative vegetation dynamics module. The answer is sort of – but there are a lot of challenges with this, as well as other coarse-grained modularity, e.g., due to the fact that it can be hard to draw a clear box around what's included in a given parameterization (e.g., which photosynthesis scheme should we use?).
  • Hillslope modell. Now includes downscaled meteorology (taking into account slope & aspect)
  • Perturbed parameter experiment
  • Integrating NEON ecological observations, enabling real-time ecological forecasting

Transition to CTSM

Needs:

  • CTSM overview paper
  • CTSM website
  • Capability to run CTSM in parallel to NoahMP for various applications
  • Funding to support assessment of CTSM

Bill Sacks raises the question: Should the complete separation of flux calculations from state updates remain a high priority? Bill's feeling is that this may have less payoff at this point than some of the other goals, such as broadening CTSM's application areas.

Regarding the paper:

  • Jean-Francois asks if it would be worth having two papers: One giving the high-level vision (e.g., for BAMS); one giving next steps and more details that would be more appropriate for GMD or JAMES.
  • Dave sees that point; it may come down to whether we have the time to push through two papers. However, also note that Rosie and Charlie have written a high-level philosophical paper that we can already lean on. (https://agupubs.onlinelibrary.wiley.com/doi/full/10.1029/2018MS001453)
  • Bill Mahoney: Couldn't you have a BAMS or other Journal paper and an NCAR Tech Note to handle the details?

Gordon sees two things: unifying models; and demonstrating that CTSM is doing well, in order to build the user community. That could argue for two separate papers.

Mariana: One important thing CTSM provides is facilitating a hierarchy of complexity. This is a big need of the community, and she feels it's worth highlighting.

Dave L: Original idea was to bring all NoahMP parameterizations into CTSM. But now, given resource limitations, he feels it may be more important to bring in major NoahMP capabilities (e.g., ability to control specific crop planting date); he isn't hearing a lot of need for many specific NoahMP parameterizations that aren't already in CTSM. (Similarly, the original idea of fully separating flux calculations from state updates may not be the highest priority, unless we feel we're getting bad answers due to the lack of such separation.)

Danica: suggests for the paper highlighting the scientific performance benefits you can get from CLM-SP, showing the additional things you can get from this tool – not just that it can be a replacement for NoahMP.

Website: Dave would like to have a CTSM website available in parallel with the paper (with links to existing websites).

New ESMF/NUOPC coupling infrastructure (CMEPS), data models (CDEPS) and impacts on CTSM (Mariana)

New coupling infrastructure, CMEPS, will be the default for CESM2.3 and beyond.

CMEPS enables both new functionality and collaborations. On collaborations, we can now share components and infrastructure between NCAR and NOAA systems. On new functionality, two that are of particular importance to CTSM / CESM are: no longer need to generate most mapping files when introducing new grids; and no longer need to generate domain files for new grids.

CDEPS: among other things, enables new on-line regridding capability between forcing data grid(s) and model grid.

New proposal recently funded in NOAA's UFS: extracting land model and coupling it via a NUOPC cap. So, even though NoahMP is the initial target, this will facilitate running CTSM as an alternative land surface scheme (along with coupling to JEDI).

  • This is probably order months away from being operational, barring unforeseen issues.
    • The hardest piece of this is probably making some changes to the atmosphere to allow it to interact with an external component. But there are multiple efforts now underway to address this.

Transition to CTSM: Funding (and other discussion)

Bill Mahoney asks: how well is CTSM already able to replace NoahMP?

  • Dave replies: There are still some needs like specifying crop planting dates, but he would cautiously say that, for regional applications, the scientific performance of CTSM seems good. Fai points out that there is a need for broader evaluation, but there isn't funding for this now.

Dave L asks what the priority should be for coupling to LIS.

  • Andy Wood points out that this opens the door to CTSM participation in NLDAS.

Aug 26

CLM / CTSM Accomplishments

Examples of fields where CESM, since its inception, had a profound impact on, changing the course of that particular field, leading to breakthrough science with major / broad impacts including on society. Ideally, CESM should be the first to recognize the importance of these efforts, leading them, and showing its breakthrough impacts.

  • For each such topic, a slide is requested stating the importance of the work and its impacts. If you can provide such a topic and its slide, please send them my way by 31 August.
  • This figure from Rosie & Charlie's paper may help spark discussion.

CTSM website

Aug 12

Announcements

  • Welcome
    • Katya Jay, postdoc coming to CU Boulder
    • Dan Muccio, graduate student at Michigan
    • Others
  • CTSM tutorial
    • May 23-27, 2022

NEON collaboration and developing single point capabilities in CTSM

July 14

Announcements

  • CTSM tutorial
    • in person, suggested May/June 2022
  • MIP and other support
    • TRENDY: Data just posted Daniel, will figure out what needs to happen here?
    • SMYLE: Steve wants to run near term forecast with CESM2, with components initialized through May 2021. Sean's started this process and we need to coordinate next steps.
    • ISIMIP and RECCAP simulations that Wim and Dev have volunteered to help with are using input data from GFDL. I'm assuming they're doing these simulations on their own machines and should generate their own input data, but maybe Sean's script can help with this? Hopefully this is one we can provide with minimal support...

CESM2-LE results

  • Will presents preliminary results on 21st century changes in snow, runoff, and growing season length

July 1

Joe Hamman presents on his work at CarbonPlan

May 20

Announcements

  • CESM Workshop. Registration still open for attendance.

Updates: FATES-SP, Rosie

  • FATES-SP mostly working with PR to FATES-main pending
  • CLM5-SP != CLM-FATES-SP: Fast gas exchange has diverged since 2012. (LUNA, Medlyn, acclimation, etc).
  • Unwinding these changes makes the models look more similar, but FATES also requires consideration of:
    • Crops
    • SAI bug
    • leaf layers & direct vs diffuse light
  • Long term, this raises questions about how we maintain the duplication / redundancies of photosynthesis code in CTSM and FATES.
    • How does the whole system work when you 'roll forward' the CTSM and FATES codes,
    • Especially when Big-Leaf is used for crops and FATES is used for natural vegetation?
    • Are we close enough to having an efficient FATES code that can replace the big leaf radiation scheme? This may require a big-leaf implementation in FATES.

Updates: Hillslope model, Sean

  • Previously implemented:
    • Slope and aspect affects on radiation has been implemented.
    • Soil thickness can also vary based on slope position, aspect, etc.
    • Elevation-based temperature downscaling
  • New elevation-based temperature downscaling
  • Remapping from 16 column hillslope simulation to high resolution shows spatial structure within each grid.
  • Now trying to maintain large scale relationships using irregular grids / mesh

May 6

Announcements

Computing resources, are there community projects we should saving this for, or consider donating hours to CESM testing?

  • 8.4 million on LMWG Prod (93300041) 10/31
  • 2.4 million on LMWG Dev (93300641) 10/31
  • 2.7 million on TSS (08010000) 9/30

Obligations and Requests

  • 1M PPE for 1-at at time
  • 1M+ PPE for latin hypercube
  • Trendy (try running 0.5 degree?) & other MIP requests
  • Meg & climate process team
  • Claire & Abby, CAM-CLM perturbed simulations
  • Jackie, tropical FATES & Fire
  • 400k Sanjiv for high res.
  • 1.5M Peter for
  • 500k Andy Newman, CTSM high resolution calibration on NNA project.
  • 500k Andy Wood, CTSM and S2S predictions (if needed)
  • Polly Boutte, regional FATES simulations
  • Andy Fox & Xuli, Global DA on LAI
  • Danica, Crops & clouds AMIP runs
  • Rosie, FATES calibration efforts (2.5 M on CSL allocation)

April 22

Katie leads a discussion on ESDS goals which likely will include development of new python-based diagnostics packages for CESM component models.

Slides are here

  • This will be a bottom-up effort.
    • Funding / people / time spring to mind as issues
    • Maybe NCAR-NEON activity will be a good place to start!

Example projects for reference

LMWG opportunities

  • CTSM-py https://github.com/NCAR/ctsm_python_gallery,
    • A place to put sample workflows and tools that use CTSM model output
  • PPE-tools - Daniel Kennedy, Katie Dagon (under development)
    • Tools to support scripting and analysis of the CLM5 perturbed parameter ensemble (PPE)
  • NCAR-NEON
  • LMWG diagnostics?

April 15

CESM Workshop, Virtually 14 - 17 June 2021

  • LMWG 3.5 hours Tues. morning
  • Overview presentation (this is expected, helpful)?
  • Topical discussions centered on:
    • Disturbance and Management
      • Fire - Charlie Koven - probabilistic vs prognostic method of handling fire and impacts
      • Agriculture
    • Interactions and Feedbacks
      • Hillslope Hydrology
      • Mizuroute?
      • FATES-SP - priorities and progress, results?
      • NEON simulations
  • Series of <5 minute lightning talks (~6 talks per topic)
  • Coordination committee
    • Jackie, Rosie, Peter, Charlie, Negin, Sanjiv, Adrianna, Stefanie
  • Invited talks only? welcome poster presentations (some of which may be elevated to main session)
    • moderated discussion with guiding questions
    • single slide introduction to posters?

Energy closure issue, Martyn Clark

April 8

Bill and Danica lead a discussion on grid cell averaging

Link to slides from this discussion

Background

  • Sparked by discussion regarding history fields related to ozone code.
  • How does grid cell averaging (from subgrid, e.g., column or patch level, especially when a variable is not calculated on some land units)?
    • extensive vs. intensive quantities (gC/m2 vs. K)
  • How do users generate regional or global averages, especially with some fraction of a grid cells has land units w/ no values.
    • Should be weighted by fraction of grid cell that a quantity is calculated over
      • This is further complicated by LULCC, since fields are updated annually.
    • We don't currently provide the metadata to do this correctly or which fields should be used for each variable of interest.
    • Maybe use pct_nat_veg, pct_crop & other land units, but additional metadata are needed on history files for this to be done correctly.

How this is done in the code.

  • LOTS of different ways...
  • l2g_scale_type provides a way to do this for veg, crop, urban, ice, etc.
    • Bill seems to like this since it clarifies metadata associated with each variable.
    • Daniel asked if this could have veg_intensive and veg_extensive definitions to handle most of the history fields.
  • set_*, can be set to spval or 0. depending on if a quantity is included in averages or not.
    • Bill doesn't seem to like this
    • Erik also think this this should not be used
  • most history fields are typically set to spval, so they are not included in gridcell averages unless they are defined elsewhere.
  • Other variables are later set to zeros later in the code, which doesn't lend itself well to documentation in metadata.
  • Peter asked how we maintain backwards compatibility in old history files and how we do our analyses?
    • Bill thought this seemed more like a bug, if averaging was done incorrectly.
  • Katie asked what other modeling group do?
    • Dave noted this would be a great topic for the land model summit, if it ever happens.

Questions to consider moving forward

Is there an agreement on a path forward?
  • We need to add metadata to document how the field was averaged, as well as provide the landuse fractions so users can do what’s needed with it.
  • l2g_scale_type seems like a good way to do this, especially for new history fields.
  • Crops & urban seem like particular use cases that may deserve additional consideration.
How do we get there, at what priority?
  • It's somewhat daunting given the 1400 fields we can write out on history files.
  • Issue seems greater w/ biogeophysics (BGC seems OK?).
  • Write out l2g_scale_type metadata on history files.
    • Bill's still concerned this may be misleading based on set_* = spval, although we're we're explicitly setting l2g_scale_type this may be OK?
  • Dave suggested picking a small set of variable to do this correctly.
  • Possibly extend this to more commonly used variables used in analyses.
  • Also need to build in the tools so that analyses are done correctly (ESDS & CTSM_py activities).

March 25

Ronny Meier introduces project on roughness lengths for deciduous vegetation [#1316]((https://github.com/ESCOMP/CTSM/issues/1316)

March 11, 2021

LMWG discussion

  • Congratulations to Yue and Inne, 2021 Slater Award winners
  • Surprises, thoughts, or recollections?
  • Considerations for CESM Workshop: Panel based discussions around themes (e.g.):
  • Discussion questions before meeting
  • Series of lightening talks from (invited?) speakers:
  • Themes:
    • natural disturbance (fire, windthrow, drought)
    • management (managed forests, agriculture)
    • land-hydrology connections
    • land-atmosphere interactions

Updates on PPE and CTSM5.1 tag(s).

  • CTSM5.1.0 tag will also rename master to main

Discussion of CTSM5.1 simulations

Crowd Sourcing model diagnostics

  • do we pass the 'eye' test Dave, for Nate and ILAMB

Web page(s): Keith.

Feb 11, 2021

Simulating DOM fluxes in CTSM, Devaraju

Updates on PPE tag & state of CTSM5.1 simulations

Planned simulations with CTSM5.1

  • Land only big leaf & FATES reduced complexity modes.
    • Reduced complexity FATES needs to be integrated PR 1182.
    • For transient LULCC we need the raw LUH2 data brought in as streams files #1077.
  • CESM-FATES smoke test (waiting for FATES-SP + CESM2.2.beta02 tag)

LMWG meeting

Jan 28, 2021

Brainstorming ideas for land contributions to CESM2-LENS

  • Brainstorming JamBoard
  • Pages are organized to show:
    • Mature work (Keith and Ryohei's work at Pusan),  
    • Topics / analyses that are (relatively) developed,
    • Idea for additional analyses,
    • Additional Questions & Requests &  
    • A quick look at some data (very preliminary).
  • Please contribute ideas and thoughts to the JamBoard & contact Keith R. if interested in pursuing a topic in greater depth.

Jan 14, 2021

Updates

  • LMWG and BGCWG meetings Feb 23-25, 2021. https://www.cesm.ucar.edu/events/wg-meetings/2021/land.html
  • Biomass heat storage PR to CTSM5.1
    • Stocking density as new CLM5.1 parameter
    • FATES integration beginning
  • Community simulations of CESM2 with forcing from CESM1-LENS runs.
    • Investigate the relative role of forcing changes versus model structural changes in the differences between CESM2 CMIP6 runs and the CESM1-LENS runs.
    • Marika and Cecile are considering putting together an overview paper that describes these runs with a few key results.
    • Interest in contributing to paper or analyzing results?
  • CESM2-LENS simulations, discussion at 4 MST Jan 28.

Isla's findings on improved wintertime temperature variability

  • Update from Fall 2020, related to snow density!

WUE diagnostics

  • Daniel introduces different flavors and how to calculate them