Climate Risk Analysis

The prime climate research interest of the CRA team collaborating in GATEWAYS is to determine the tele-connections between the regional surface ocean climatology and key climate oscillators (WP1, Task 2). This requires (1) excellent data (measurement, proxy, timing) and (2) state-of-the art methods of statistical time series analysis. (1) Data are obtained via the GATEWAYS partners. (2) Statistical methods are developed and adapted by CRA on the basis of existing own software. The tele-connection estimations serve the other project partners in the climatological interpretation. The developed software is made available to the project partners and has a lifetime beyond the closure of the project. This means a long-term impact on climatological research.

The data produced by the GATEWAYS partners contain information about past climates comes from proxy measurements and climate models. Statistical processing of these data focuses on (1) the Lomb-Scargle periodogram for bivariate spectral analysis and (2) Pearson's and Spearman's coefficients for correlation analyses. Both methods have two input time series, one representing the Agulha Current, the other representing a potential climate driver. The project-specific adaptation of the Lomb-Scargle method consists in taking into account dating uncertainties (e.g., of sediment cores). The project-specific adaptation of Pearson's and Spearman's correlation coefficients consists in taking into account non-normal data distributions and autocorrelation effects. Both adaptations require extensive mathematical simulations and, hence, considerable computing resources. The general motivation for taking into account the real-word climate data properties (dating error, non-normal distribution, autocorrelation) is to provide robust uncertainty estimates with the result. This prevents from making overstatements. It is planned to implement the adapted methodologies as Fortran 90 programs, allowing for large data sizes and efficient calculations. It is further planned to include a graphical user interface to facilitate interactive working. The new software is called REDFITMC3 on Lomb-Scargle and CORR-PS on correlation. The software is applied to the tele-connection data sets and the results interpreted in close cooperation with the GATEWAYS partners. The programs, with some of the theory behind it, shall be presented to the partners at an internal workshop.

VUA is an important collaboration partner: literature, data, interpretation, user-input into the development of the software. VUA overtakes WP1 Task 1 (estabilishing the surface-ocean data). Also the other partners (1 to 4, 6 to 8) provide their (paleo-) climatological knowledge, and data, and help with the interpretation of the statistical results. In turn, it is expected that all GATEWAYS partners shall benefit from the results, and the availability of the method, of statistical time series analysis.

Partner Profile

Climate Risk Analysis (CRA) is a research company working on risk quantification of extreme climate/weather events and statistical analysis of climate time series. CRA is young and expanding, it was founded in 2005 by Manfred Mudelsee and resides in Hannover, Germany. CRA has pioneered mathematical simulation approaches for realistic quantification of the uncertainties of estimated climate parameters. CRA/Mudelsee are rooted in mathematics, physics and geology and have developed close collaborations with various Earth Science disciplines such as Oceanography or Glaciology. See for more information about CRA.

Visit Homepage of Climate Risk Analysis