Dirk Wolters

We have applied quantile regression forests (QRFs) to generate probabilistic forecasts of weather conditions associated with low-visibility procedures (LVPs) at Schiphol Airport (Amsterdam, the Netherlands). LVPs are determined by combined thresholds of cloud-base height and (runway) visibility. Forecasts of these conditions are critical for airport operations, as they inform operational planning, with the potential of minimizing meteorologically induced disruptions. Using a dataset of 5 years of hourly data, we have performed a forward feature selection and optimized QRF’s hyperparameters for this specific application, and evaluated the model’s performance for different forecast lead times and different LVP classes. Hereby, LVP forecasts were obtained by combining separate models for cloud-base height and (runway) visibility, applying a Schaake shuffle approach for restoration of the dependencies between these parameters. The verification revealed consistent positive Brier skill scores (BSSs) for the three most common LVP classes: marginal, A, and B. Although the skill was not always positive for the more extreme LVP classes, C and D, we argue that also for these conditions forecasters might derive valuable indications from the forecast system. We demonstrate the operational utility of the system with an example, also illustrating the support of interpretability through the use of Shapley additive explanation (SHAP) values. Our results underscore the potential of QRF for probabilistic forecasts of meteorological conditions, for aviation and other purposes.

A better quantification of the urban heat islands (UHIs) in the Netherlands is urgently needed given the heat stress–related problems in the recent past combined with the expected temperature rise for the coming decades. Professional temperature observations in Dutch urban areas are scarce, however. Therefore, this research explores the use of observations from weather stations that were installed and maintained by weather amateurs. From a set of over 200 stations, suitable and representative data have been selected from 20 stations, using a set of objective selection criteria that are based on metadata. One year of data (January–December 2010) was considered. From these data, estimates have been obtained of the magnitude of the UHI in Dutch low-rise residential areas. A positive relation (linear model with r² ≈ 0.7) was derived between the summer-averaged UHI and the (neighborhood scale) population density around the observational sites. It was found that the UHI in summer is strongest in nighttime conditions and that it increases with decreasing wind speed, decreasing cloud cover, and increasing sea level air pressure. The summer-averaged UHI was ~0.9°C. During nighttime in a relatively warm 1-month subperiod of the summer, the average UHI was ~1.4°C. During spring and autumn, the UHI was lower than in summer; during winter, no significant UHI was observed. The agreement in results among the different stations and the accordance of the magnitude and variation of the observed UHI with those described in the literature show that automatic observations from weather amateurs can be of sufficient quality for atmospheric research, provided that detailed metadata are available.