Timeseries development

[gavinny]

The is for discussions related to the merging of data sources to create the input files and assessments of uncertainties.

[durack1]

@gavinny looking forward to this discussion, much to learn and calibrate!

[AndyShepCPOM]

IMBIE

Spatial Resolution
The IMBIE team will be increasing the spatial resolution of our reconciled ice sheet mass balance assessment to drainage basin scale over the coming months and we are very happy to contribute this dataset to the project.

Temporal resolution
Our dataset will run from ~1980 to present day and will be annual in temporal resolution. If finer is required/desired please let us know

Floating ice
We would be interested to consider extending our assessment to include floating ice (ice shelves & ice tongues, not sea ice). It seems that there are a variety of datasets currently in use (Depoorter, Rignot, Adusimilli, Slater, Greene, etc) and so (similar to land ice mass balance) a reconciled estimate could be of value to the user community.

[AndyShepCPOM]

We have several input output (IO) datasets submitted to IMBIE and so it can be computed from those, but we also have surface mass balance (SMB) and so it can be computed from altimetry or gravimetry total mass balance using that (analagous to computing mass balance from IO method, which is commonly done).

[jlbamber]

Hi Torge, for the GrIS we can do a lot better than assume balance prior to 1970. See e.g. Box and Colgan 2013, which has been updated by Ken to produce a total mass balance (TMB) time series for 1840-next week. I mention this in my talk. It comes with uncertainties. Of course, it's not perfect but much better than assuming balance esp for the 1930-40s when it definitely wasn't.

For the AIS it's a bit more tricky but, as I suggested in my talk, Frederiske (2020) manages to achieve SLB closure for 1900-20XX with an assumption for AIS mass trends that are pretty close to balance and I think this is pretty compelling for earlier 20th C. Jonathan Gregory did the same in an earlier study but Thomas's is a more robust reconstruction.

[mankoff]

@AndyShepCPOM

We have several input output (IO) datasets submitted to IMBIE and so it can be computed from those

To be pedantic, I/O as our community uses the term isn't sufficient. You need the I to be provided with constituent terms (snowfall, rain, runoff). If the IO provided to you is just net SMB & discharge, you've lost necessary information.

[AndyShepCPOM]

IMBIE has a surface mass balance experiment group who assess regional climate models and their various elements, including net surface mass balance

[mankoff]

Discussions on historical balance are ongoing on a different thread: #13

[durack1]

Looks like there's already a heap to collate here! For the CMIP7 DECK/CMIP historical experiment, we expect to collate forcings that cover the 1850-2025 (or the latest available) period. Providing time histories that extend back that far would be ideal if it were possible. Uncertainties will be huge (or signals will be tiny)... This is also true of other forcing datasets.

For a little more context, we have a CMIP Forcing Task Team (TT) homepage at https://wcrp-cmip.org/cmip7-task-teams/forcings/, and the current timelines that we are hoping to hit are located at https://zenodo.org/records/10326376 (and have been updated 7 times already). We have started to see the first batch of new prototype forcings land for review already. In brief, we expect to collate the first complete set of CMIP:DECK:historical forcings by April 2023 (covering the 1850-2021 time period), and through iteration and testing, update these late 2024 and 2025 accounting for feedback, and attempting to temporally extend these as far forward to near real-time as possible. The aim then is for their use in the CMIP7:DECK:historical (plus esm-historical, *piControl, etc)

[durack1]

I added a comment in discussion #13 regarding how I have been imagining the process (of timeseries generation, to model testing, to proposing this as a CMIP:DECK:historical experiment amendment) would occur - see #13 (reply in thread).

I hope this is helpful, happy to hear alternative views that improve on these initial thoughts

[jlbamber]

So far not been much discussion of GIC so here are a few points that came up. As for the GrIS global mass balance estimates existing for the GRACE period and reconstructions, combining glaciological, geodetic and statistical methods extend back to 1960s. Reconstructions based on re-analysis exist for earlier part of the 20th C (e.g. Marzeion et al 2012) with larger uncertainties. What proportion of the GIC imbalance enters the ocean as a FW anomaly is not well constrained for some of the RGI sectors, the most important of which is the HMA. However, these flux anoamlies (for individual RGI regions) are small compared to the ice sheets and other FW flux sources and sinks. The exception to thiss, and where the FW anomaly may be important ofr ocean processes, is in the Arctic. Over half the GIC volume lies in the Arctic and here the assumption that it enters the ocean as an anomaly is reasonable to first order. A reconstruction for 1950-2021 is available for Arctic GIC. Prior to this will require using statistical approaches such as Marzeion 2012.