Climate Change Information Distillation for Decision Scale Applications
Author: Yinpeng Li
Posted: Friday, August 19, 2016
Over the last several years the international climate modelling community has put a great effort into the dynamic downscaling realm with now massive regional climate data sets available on the internet. For example, ESGF@LiU/CORDEX (https://esg-dn1.nsc.liu.se/search/cordex/), users can download CORDEX raw data.
The CORDEX vision is to advance and coordinate the science and application of regional climate downscaling through global partnerships.
CORDEX Goals
- To better understand relevant regional/local climate phenomena, their variability and changes, through downscaling.
- To evaluate and improve regional climate downscaling models and techniques.
- To produce coordinated sets of regional downscaled projections worldwide.
- To foster communication and knowledge exchange with users of regional climate information.
CORDEX data is good for the analysis of the fundamental processes and dynamics of the climate system on multiple scales, as reflected in the peer-reviewed publications, yet the processed data remains opaque and mostly inaccessible to external/end user communities. As Professor Bruce Hewitson commented in a CORDEX 2016 workshop early this year the information distillation for end users still poses a dilemma for scientists.
With many issues regarding quality and interpretation, CORDEX data is still perceived as the primary data source for legitimate climate change risk assessment and resultant adaptation actions. The team at CLIMsystems has therefore taken the step of processing the growing volume of CORDEX RCM data for applications just as we do GCM and other data sources.

Needs-driven analysis: usually somewhat simplistic post-processing of data sets (bias correction, mean and range) attempts to address the (perceived) needs of end users. The goal being to provide tailored information through the lens of system complexities at the decision-scales of thresholds and risk. The range of GCM and RCM data therefore attempts to span the breadth of possibilities and likelihood of outcomes.

Needs-driven analysis: usually somewhat simplistic post-processing of data sets (bias correction, mean and range) attempts to address the (perceived) needs of end users. The goal being to provide tailored information through the lens of system complexities at the decision-scales of thresholds and risk. The range of GCM and RCM data therefore attempts to span the breadth of possibilities and likelihood of outcomes.


CORDEX Africa for Ghana applications


Climate change case specific analysis vs online data and tools
Online tools related to climate change data are becoming more and more common. While it is good for awareness raising and capacity building, several issues need to be considered by online data tool users:
(1) Climate variables: the majority of the decision support and adaptation support online data portals display mean monthly temperature and precipitation data, or bioclimatic variables, which are not directly linked to risk management and risk reduction, or governance, where climate extremes and derived variables are more important, including for example: drought and flood, extreme precipitation, extreme wind, heat stress, dust haze and sand storms and storm surge. All these extreme variables need dedicated site specific analysis in order to make the results relevant enough for either planning or risk governance.
(2) Data quality control and cross validation: The quality of climate data needs to be carefully checked, and cross validated with other data sources. Directly applying one online data source could be dangerous. Many sources could produce the wrong climate data: observation equipment faults, recording errors, digitising errors, model configurations, mislabelling, and typos are just a few of the examples where data processing can go wrong. Without careful specific quality control (and documentation expressing where the underlying data came from and how the data was processed), end users should be careful.
(3) The newness and completeness of the data: Online data tools usually are developed through a project, when the project is finished, the data will stop being updated. With progress in climate science, more and more scientific raw data becomes available, but may not be easily accessible to non-climate professionals. Therefore most of the online data tool are partial data at best, which could cause significant bias in historical and future projection analysis. In contrast, case specific analysis carried out by ethical climate service providers apply the newest and complete datasets for analysis and maintain scientific credibility.
(4) Data interpretation: Climate data should be interpreted by professional climate experts, who have sufficient climatological knowledge and scientific credibility to understanding the meaning of the data. Misinterpretation of the data is a big risk of using online information without good understanding of the underlying climate change science.
(5) Scientific credibility: Online tool data tends to simplify the data and information, in order to make it easy to access, however, this can lead to the loss of scientific soundness. The main purpose of online tool embedded data is for awareness raising or capacity building, but it is often less robust for real operational decision making and risk management and governance, because of the reasons listed above.
(6) Cost effectiveness of case specific analysis reports: It seems in strategic planning stages, people only need first order assessments, or rough analysis, however, when a company really tries to develop a robust strategic plan, it will be realised that the relevancy and specification of data is important information. Lacking solid scientific soundness, a strategy or plan cannot go very far, and the strategy or plan may need to be revisited which can be costly. Therefore, case specific climate change analysis is often necessary and cost effective for clients who are facing real climate risks in their business.
Table 1: CORDEX Data Availability for the Africa Domain
The code 6 under ‘Variables’ for the following tables includes: precipitation, mean temperature, maximum temperature, minimum temperature, solar radiation, wind speed. The code 7 under ‘Variables’ indicates means the above six variables plus relative humidity.
ID | GCM RCM combination | Variables |
1 | AFR-CanESM2-CRCM5 | 6 |
2 | AFR-CanESM2-RCA4 | 7 |
3 | AFR-CNRM-CM5-CCLM4-8-17 | 6 |
4 | AFR-CNRM-CM5-RCA4 | 7 |
5 | AFR-CSIRO-Mk3-6-0-RCA4 | 7 |
6 | AFR-EC-EARTH-CCLM4-8-17 | 6 |
7 | AFR-EC-EARTH-HIRHAM5 | 7 |
8 | AFR-EC-EARTH-RACMO22T | 7 |
9 | AFR-EC-EARTH-RCA4 | 7 |
10 | AFR-EC-EARTH-REMO2009 | 6 |
11 | AFR-GFDL-ESM2M-RCA4 | 7 |
12 | AFR-HADGEM2-ES-CCLM4-8-17 | 6 |
13 | AFR-HADGEM2-ES-RCA4 | 7 |
14 | AFR-IPSL-CM5A-LR-REMO2009 | 6 |
15 | AFR-IPSL-CM5A-MR-RCA4 | 7 |
16 | AFR-MIROC5-RCA4 | 7 |
17 | AFR-MPI_ESM-LR-CRCM5 | 6 |
18 | AFR-MPI-ESM-LR-CCLM4-8-17 | 6 |
19 | AFR-MPI-ESM-LR-RCA4 | 7 |
20 | AFR-MPI-ESM-LR-REMO2009 | 6 |
21 | AFR-NorESM1-M-HIRHAM5 | 7 |
22 | AFR-NORESM1-M-RCA4 | 7 |
Table 2: CORDEX Data Available for the Middle East and North Africa (MENA) Domain
ID | GCM RCM combination | Variables |
1 | MNA-22-EC-EARTH-RCA4 | 7 |
2 | MNA-22-GFDL-ESM2M-RCA4 | 7 |
3 | MNA-44-CNRM-CM5-RCA4 | 7 |
4 | MNA-44-EC-EARTH-RCA4 | 7 |
5 | MNA-44-GFDL-ESM2M-RCA4 | 7 |
Table 3: CORDEX Data Available for Central America Domain
ID | GCM RCM combination | Variables |
1 | CAM-44-EC-EARTH-RCA4 | 7 |
2 | CAM-44-HADGEM2-ES-RCA3 | 7 |
3 | CAM-44-MPI-ESM-LR-RCA4 | 7 |
Table 4: CORDEX Data Available for the South Asia (WAS) Domain
ID | GCM RCM combination | Variables |
1 | WAS-44-CNRM-CM5-RCA4 | 7 |
2 | WAS-44-EC-EARTH-RCA4 | 7 |
3 | WAS-44-GFDL-ESM2M-RCA4 | 7 |
4 | WAS-44-IPSL-CM5A-MR-RCA4 | 7 |
5 | WAS-44-MIROC5-RCA4 | 7 |
6 | WAS-44-MPI-ESM-LR-RCA4 | 7 |
7 | WAS-44-MPI-ESM-LR-REMO2009 | 6 |