A repeatable and reproducible modelling workflow using the ...

A repeatable and reproducible modelling workflow using the Vegetation Optimality Model and RENKU

R.C. Nijzink1, C. Ramakrishnan2, R. Roskar2, S. Schymanski1

1 Luxembourg Institute of Science and Technology, Belvaux, Luxembourg, 4Swiss Data Science Center, Zurich, Switzerland

Conclusions

Introduction Motivation

Renku

Case study Description

Data and analyses

Renku Getting started... Datasets Tracking files

Reproducing results Knowledge graph

Supported by the Luxembourg National Research Fund (FNR) ATTRACT programme (A16/SR/11254288)

Next

MOTIVATION

InIntrtorodduucctitoionn CCaasseesstutuddyy RReennkkuu CCoonncclulussioionnss

Previous Home

2

2

Next

More code openly available!

More open source datasets!

?

?

?

More computational power!

Large-scale modelling Data assimilation

Sensitivity analyses

Model inter-comparisons

Model complexity

?

Larger groups of

collaborators!

?

?

How to ensure reproducibility and repeatability?

Scientific result

RENKU RENKU

Environment for collaborative, reproducible data science

InIntrtorodduucctitoionn CCaasseesstutuddyy RReennkkuu CCoonncclulussioionnss

Previous Home

3

3

Next

Concepts

? Tracking of scientific steps to create data lineage, i.e. a knowledge graph

? Updating of out-dated results ? Tool to re-use or re-run analyses ? Sharing of analyses

Usage

? Command Line Interface ? Web platform

Features

RENKU is based on : ? Gitlab ? JupyterHub ? Kubernetes ? Keycloak ? Common Workflow Language

More info

InIntrtorodduucctitoionn CCaasseesstutuddyy RReennkkuu CCoonncclulussioionnss

Previous Home

4

4

Next

CASE STUDY - DESCRIPTION

Study sites

Six sites with a strong precipitation gradient along the North Australian Tropical Transect.

Vegetation Optimality Model Optimizes vegetation properties to maximize the Net Carbon Profit, i.e. the difference of carbon uptake by photosynthesis and carbon costs of the system. Research Question Does maximization of Net Carbon Profit explain vegetation behaviour in savanna sites along a precipitation gradient?

See session HS10.1 General Ecohydrology, Thursday 7th of May 8.30 ? 10.15 for the scientific results!

InIntrtorodduucctitoionn CCaasseesstutuddyy RReennkkuu CCoonncclulussioionnss

Previous Home

5

5

Next

CASE STUDY ? DATA AND ANALYSES

? Repeatability important: ? New insights ? Bugs

? ``Failures'' should not be be forgotten!

Never bugfree!

Always under development!

Never bugfree!

Always newer versions available!

Input data

Meteorological data SILO DataDrill Atmospheric CO2-levels Mauna Loa

Soil and Landscape Grid of Australia

Analyses

Python-scripts pre-processing Shell-scripts

VOM-code

Python-scripts post-processing

Never satisfactory at once!

Final result

Validation data

Flux tower CO2-assimilation

Flux tower evaporation

Fpar-based vegetation cover

 ................
................

In order to avoid copyright disputes, this page is only a partial summary.

Google Online Preview   Download