Teleconference on land-hist papers, July 17, 2020

Participants: Sonia Seneviratne, Hyungjun Kim, Lawrence Mudryk, Agnès Ducharne, Andrea Alessandri, Binghao Jia, Daniele Peano, Franco Catalano, Frédérique Chéruy, Jiafu Mao, Rachael Turton, Ryan Padrón, Stefano Materia, Nicolas Vuichard, Wilhelm May, Yukihiko Onuma, Gerhard Krinner, Eleanor Burke

1. Update on ETH data server

Selected monthly variables have already been downloaded at ETH by Ryan. Further monthly variables interesting for all 3 studies are now being downloaded at ETH by U. Beyerle. Daily data have not yet been downloaded, but could be downloaded if necessary (disk space is a obvious limit). Lawrence was able to access the data. Hyungjun also archived some data himself and thinks he can access data from the ETH repository. For those who want to use this repository, please contact Urs Beyerle (

2. Hyungjun Kim Part I

Hyungjun presented some comparisons of Land-hist vs CMIP and AMIP, based on 3 models: CESM2, IPSL, MIROC6.

Comparisons with GSWP3-Fluxcom dataset from Jung et al. 2018 (ET, net radiation, sensible heat flux, P-ET)

For precip: Similar performance of CMIP and AMIP LS3MIP? simulations.

Can compare relative performance for different variables, e.g. radiation vs P-ET. Performance of LMIP is better for P-E than radiation.

Some comparisons of SM-LH coupling using ESA-CCi and Fluxcom for evaluation. Analyses will be expanded to other GCMs once other models will be downloaded from the ETH server. So far, differences between models do not seem related to L-A coupling.

Discussion -

Jiafu Mao: Consider observational uncertainties? Other soil moisture and ET products could be considered. Hyungjun: Currently no plans to include more soil moisture datasets.

Sonia: If no more datasets are included, some error bars on observations datasets would be useful.

Discussion between Hyungjun, Agnes and Stefano about the possible inclusion of other L-A coupling metrics. Possible inclusion of more analyses needs to be discussed later. Frederique C.héruy has developed other measures of coupling metrics using histograms

General remark: Clear links between the three papers will have to be made.

Inclusion of ILAMB metrics to be defined

3. Ryan Padrón: Part II

Focus of this paper is on land water and carbon variables. Ryan need names of model contributors, particularly from modeling groups and contributors of specific analyses, for work on this paper.

Input from modeling centers is required on Table describing the models' characteristics with respect to carbon cycle and soil moisture, and representation of vegetation.

Some planned analyses are:

Dave: WUE

Andrea: active rooting depth; ET/P

Main variables: ET, soil moisture (upper, root-zone, total), carbon.

Ryan presents some preliminary analyses: Timeseries of land water balance components

There are small differences in precip between models (this was also found by Hyungjun)

Large differences in ET/P between models (about 15%)

Comparisons with precipitation observations are currently only done with GSWP3

Ratio of transpiration / ET shows interesting results

Analyses of carbon balance components:

Some interesting features: LAI is different across models, why?

For land use, the raw data is all from LUH, but the interpretation may be different between models

Analyses of long-term changes:

1985-2014 vs 1901-1930: Substantial changes, but also substantial intermodal differences.

Further planned analyses:

comparisons with ISIMIP and Trendy

extreme years

attribution to human-induced climate change


Several models not yet CMORized (UKESM+HadGEM, ELM/DOE, H-TESSEL). It might be manageable to wait for the official cmorization of the missing models if < 1 month. Many people on vacation...

CMCC: missing some water balance variables

J. Mao: Any observations to be included in the paper? -> Yes, some preliminary analyses. Need to check which data are available for carbon.

Agnes: Is there a significant trend in precip between early 20th century and late 20th/early 21st that explain some of the land trends? -> To be checked

Ryan will consider if including CMIP and AMIP data could be interesting for the paper.

Hyungjun: Did not show results for trends, but benchmarking scope from Part I paper includes trend.

Sonia: include ILAMB metrics?

4. Lawrence Mudryk: Part III

Eleanor Burke presented some preliminary permafrost analyses, building on accepted paper on permafrost in CMIP6 coupled models in The Cryosphere

The paper on coupled simulations provided analyses that removed coupled model biases, focusing on soil processes. Here, the idea is to compare more directly outputs from hist (+AMIP?) and land-hist. Might consider looking at trends.

Identified observational data sets, over 20th century.

Special attention to snow-permafrost relationship, using snow insulation diagnostics. This provides a clear link to the snow section.

Lawrence Mudryk: Snow analyses

Has been thinking of ways to evaluate differences between land-hist + coupled. Nothing is currently in ILAMB. Lawrence has prepared diagnostics of integrated (hemispheric) scores (bias, RMSE, variability).

Started downloading land-hist from ETHZ

Plans to look into regional information, trends


Andrea and Franco: relate skill metrics to albedo performance (linked to SAF) - >No albedo skill metric in ILAMB yet. Meeting with Dave Lawrence & Chad Thackeray planned on ILAMB integration

No true surprises yet from historical runs... no model "good enough"...

Carbon cycle aspects? (Link to Ryan's analysis) - Currently not planned, but link exists through permafrost. However, only very few models with permafrost carbon. -> One possibility is to include specific high-latitude carbon cycle diagnostic in paper II

Eleanor: Relationship between snow depth and soil temperature isn't the same in historical and land-hist (??) - needs to be understood, needs input from modeling groups on this.

Link to 3 papers: Part I Part II: Part III:

Data requests of 3 papers:

Action items:

ALL: Have serious look at the 3 documents, provide input, volounteer for analyses

ALL: Think about links between the three papers

Modeling groups: Keep project up to date on available simulations

Next teleconference (after vacation season): Friday August 21, 1400 CEST spacer