Statistical Toolkit

Statistical methods form the necessary scientific basis for assessing trends in climate, seismicity, and the likelihood of extreme events and associated hazards.

Introduction
Traditional methods used for most KNMI products and information services assume a stationary geophysical environment. These methods seem no longer prudent in times of climate change and increased induced seismicity. The suite of statistical methods that is used at KNMI for our key products (including the climate atlas, climate scenarios and seismic hazard assessments) is due to be renewed.

Reproducibility and Reusability
KNMI should continuously improve the statistical methods used in the research process. An important factor to achieve this is to emphasize reproducibility and reusability of new methods. This implies a new way of working, for instance by incorporating a version control system not only in the development of the methods but in the whole research process. In this way different approaches can be compared much faster, which leads to new insights and incremental progress.

Current sub-projects

Fog detection from camera images (with the KNMI DataLab)
Fog is a serious threat to public safety, in particular for the transport sector. Unfortunately, fog formation can be very local and of short duration. While the local scale makes it difficult to observe consistently with the sparse network of visibility sensors operated by KNMI, the short duration hinders the coverage via satellites.
The idea of this project is to unearth the potential of image data, e.g. provided by traffic cams, to increase the observation network. With such an improved observation basis, nowcasting and the validation of the KNMI weather models can be improved.

Most of the algorithms used so far can be found here: visDec

Extreme temperatures for KNMI forecasts
The KNMI forecasts pluim is being enriched by climate information. In this project we produced return levels of daily temperatures, that take the annual cycle into account. The corresponding R-package is here: knmipluim

Monotone trend in the GPD scale parameter (with TU Delft)
In this project we developed an algorithm to efficiently estimate a monotone trend in the scale parameter of the Generalized Pareto distribution, which is then applied to the very long Central England temperature series. The corresponding R-package is here: gpdIcm

Peak values of the Central England temperature series.
Peak values of the Central England temperature series. The blue lines correspond to linear 0.5, 0.75, and 0.975 regression quantiles. The red lines are these quantiles from the monotone scale GPD approach. The dashed red line shows the 100-y return level.
Trend in mean annual temperature for selected cities in Europe, based on the EOBS data set. An interesting feature is that only the Dutch cities show a distinction between mean and median.
Trend in mean annual temperature for selected cities in Europe, based on the EOBS data set. An interesting feature is that only the Dutch cities show a distinction between mean (red) and median (blue).

Generalized additive modelling for mean temperatures
For the modelling of trends in extreme mean temperatures, e.g. the extraordinary warm September 2016 or the very mild November 2015, we are doing research on the use of generalized additive modelling.

KNMI transformation package
Scientifically not new, but a necessary step to improve the reproducibility and reusability is to put the transformation routines for the climate scenarios on solid feet.

Poisson tracking of earthquake activity (with University of Amsterdam)
We are interested in Poisson tracking of the earthquake activity in Groningen, which is more flexible than classical regression approaches.