Search Results

You are looking at 1 - 10 of 15 items for

  • Author or Editor: Maurice Schmeits x
  • All content x
Clear All Modify Search
Kirien Whan and Maurice Schmeits

Abstract

Probabilistic forecasts, which communicate forecast uncertainties, enable users to make better weather-based decisions. Using precipitation and numerous instability indices from the deterministic model HARMONIE–AROME (HA; a nonhydrostatic numerical weather prediction model) as potential predictors, we generate summer areal probabilistic maximum hourly precipitation forecasts across 11 regions of the Netherlands. We compare the skill of three statistical postprocessing methods: an extended logistic regression (ELR), a zero-adjusted gamma distribution (ZAGA), and a machine learning-based method, quantile regression forests (QRF). Forecast skill for low and moderate precipitation thresholds increases with the inclusion of extra predictors, in addition to HA precipitation. HA precipitation is the most important predictor at all lead times in ELR and QRF, while in ZAGA, the most important predictor for the location parameter shifts over lead times from HA precipitation to indices of atmospheric instability. All three methods improve upon a climatological forecast for low and moderate precipitation thresholds. ZAGA and QRF are generally the most skillful methods at moderate thresholds. QRF tends to be the most skillful method at higher thresholds, particularly during the afternoon period. Forecasts are reliable at low and moderate thresholds but tend to be overconfident at higher thresholds. QRF and ZAGA have more potential economic value than the deterministic forecast, with value remaining at high thresholds. A maximum local hourly precipitation threshold of 30 mm h−1 (a criterion in the Royal Netherlands Meteorological Institute’s code yellow warning for severe thunderstorms) is skillfully forecast by QRF in the afternoon period at short lead times.

Full access
Maurice J. Schmeits and Henk A. Dijkstra

Abstract

For a long time, observations have indicated that the Kuroshio in the North Pacific Ocean displays bimodal meandering behavior off the southern coast of Japan. For the Gulf Stream in the North Atlantic Ocean, weakly and strongly deflected paths near the coast of South Carolina have been observed. This suggests that bimodal behavior may occur in the Gulf Stream as well, although less pronounced than in the Kuroshio. Evidence from a high-resolution ocean general circulation model (OGCM) and intermediate complexity models is given to support the hypothesis that multiple mean paths of both the Kuroshio and the Gulf Stream are dynamically possible. These paths are found as multiple steady states in an intermediate complexity shallow-water model. In the OGCM, transitions between similar mean paths are found, with the patterns having similarity to the ones in observations as well. To study whether atmospheric noise can induce transitions between the multiple steady states, a stochastic component is added to the annual mean wind stress forcing in the intermediate complexity model and differences between the transition behavior in the Gulf Stream and Kuroshio are considered.

Full access
Chiem van Straaten, Kirien Whan, and Maurice Schmeits

Abstract

A comparison of statistical postprocessing methods is performed for high-resolution precipitation forecasts. We keep hydrological end users in mind and thus require that the systematic errors of probabilistic forecasts are corrected and that they show a realistic high-dimensional spatial structure. The most skillful forecasts of 3-h accumulated precipitation in 3 × 3 km2 grid cells covering the land surface of the Netherlands were made with a nonparametric method [quantile regression forests (QRF)]. A parametric alternative [zero-adjusted gamma distribution (ZAGA)] corrected the precipitation forecasts of the short-range Grand Limited Area Model Ensemble Prediction System (GLAMEPS) up to +60 h less well, particularly at high quantiles, as verified against calibrated precipitation radar observations. For the subsequent multivariate restructuring, three empirical methods, namely, ensemble copula coupling (ECC), the Schaake shuffle (SSh), and a recent minimum-divergence sophistication of the Schaake shuffle (MDSSh), were tested and verified using both the multivariate variogram skill score (VSS) and the continuous ranked probability score (CRPS), the latter after aggregating the forecasts spatially. ECC and MDSSh were more skillful than SSh in terms of the CRPS and the VSS. ECC performed somewhat worse than MDSSh for summer afternoon and evening periods, probably due to the worse representation of deep convection in the hydrostatic GLAMEPS compared to reality. Overall, the high-resolution postprocessing comparison shows that skill for local precipitation amounts improves up to the 98th percentile in both the summer and winter season and that the high-dimensional joint distribution can successfully be restructured. Forecasting products like this enable multiple end users to derive their own desired aggregations.

Full access
Maurice J. Schmeits and Kees J. Kok

Abstract

Using a 20-yr ECMWF ensemble reforecast dataset of total precipitation and a 20-yr dataset of a dense precipitation observation network in the Netherlands, a comparison is made between the raw ensemble output, Bayesian model averaging (BMA), and extended logistic regression (LR). A previous study indicated that BMA and conventional LR are successful in calibrating multimodel ensemble forecasts of precipitation for a single forecast projection. However, a more elaborate comparison between these methods has not yet been made. This study compares the raw ensemble output, BMA, and extended LR for single-model ensemble reforecasts of precipitation; namely, from the ECMWF ensemble prediction system (EPS). The raw EPS output turns out to be generally well calibrated up to 6 forecast days, if compared to the area-mean 24-h precipitation sum. Surprisingly, BMA is less skillful than the raw EPS output from forecast day 3 onward. This is due to the bias correction in BMA, which applies model output statistics to individual ensemble members. As a result, the spread of the bias-corrected ensemble members is decreased, especially for the longer forecast projections. Here, an additive bias correction is applied instead and the equation for the probability of precipitation in BMA is also changed. These modifications to BMA are referred to as “modified BMA” and lead to a significant improvement in the skill of BMA for the longer projections. If the area-maximum 24-h precipitation sum is used as a predictand, both modified BMA and extended LR improve the raw EPS output significantly for the first 5 forecast days. However, the difference in skill between modified BMA and extended LR does not seem to be statistically significant. Yet, extended LR might be preferred, because incorporating predictors that are different from the predictand is straightforward, in contrast to BMA.

Full access
Maurice J. Schmeits and Henk A. Dijkstra

Abstract

Using nonseasonal altimeter data and SST observations of the North Atlantic, and more specifically the Gulf Stream region, dominant patterns of variability are determined using multivariate time series analyses. A statistically significant propagating mode of variability with a timescale close to 9 months is found, the latter timescale corresponding to dominant variability found in earlier studies. In addition, output from a high resolution simulation of the Parallel Ocean Climate Model (POCM) is analyzed, which also displays variability on a timescale of 9 months, although not statistically significant at the 95% confidence level. The vertical structure of this 9-month mode turns out to be approximately equivalent barotropic. Following the idea that this mode is due to internal ocean dynamics, steady flow patterns and their instabilities are determined within a barotropic ocean model of the North Atlantic using techniques of numerical bifurcation theory. Within this model, there appear to be two different mean flow paths of the Gulf Stream, both of which become unstable to oscillatory modes. For reasonable values of the parameters, an oscillatory instability having a timescale of 9 months is found. The connection between results from the bifurcation analysis, from the analysis of the observations, and from the analysis of the POCM output is explored in more detail and leads to the conjecture that the 9-month variability is related to a barotropic instability of the wind-driven gyres.

Full access
Dian Nur Ratri, Kirien Whan, and Maurice Schmeits

Abstract

Dynamical seasonal forecasts are afflicted with biases, including seasonal ensemble precipitation forecasts from the new ECMWF seasonal forecast system 5 (SEAS5). In this study, biases have been corrected using empirical quantile mapping (EQM) bias correction (BC). We bias correct SEAS5 24-h rainfall accumulations at seven monthly lead times over the period 1981–2010 in Java, Indonesia. For the observations, we have used a new high-resolution (0.25°) land-only gridded rainfall dataset [Southeast Asia observations (SA-OBS)]. A comparative verification of both raw and bias-corrected reforecasts is performed using several verification metrics. In this verification, the daily rainfall data were aggregated to monthly accumulated rainfall. We focus on July, August, and September because these are agriculturally important months; if the rainfall accumulation exceeds 100 mm, farmers may decide to grow a third rice crop. For these months, the first 2-month lead times show improved and mostly positive continuous ranked probability skill scores after BC. According to the Brier skill score (BSS), the BC reforecasts improve upon the raw reforecasts for the lower precipitation thresholds at the 1-month lead time. Reliability diagrams show that the BC reforecasts have good reliability for events exceeding the agriculturally relevant 100-mm threshold. A cost/loss analysis, comparing the potential economic value of the raw and BC reforecasts for this same threshold, shows that the value of the BC reforecasts is larger than that of the raw ones, and that the BC reforecasts have value for a wider range of users at 1- to 7-month lead times.

Full access
Sem Vijverberg, Maurice Schmeits, Karin van der Wiel, and Dim Coumou

Abstract

Extreme summer temperatures can cause severe societal impacts. Early warnings can aid societal preparedness, but reliable forecasts for extreme temperatures at subseasonal-to-seasonal (S2S) time scales are still missing. Earlier work showed that specific sea surface temperature (SST) patterns over the northern Pacific Ocean are precursors of high temperature events in the eastern United States, which might provide skillful forecasts at long leads (~50 days). However, the verification was based on a single skill metric, and a probabilistic forecast was missing. Here, we introduce a novel algorithm that objectively extracts robust precursors from SST linked to a binary target variable. When applied to reanalysis (ERA-5) and climate model data (EC-Earth), we identify robust precursors with the clearest links over the North Pacific. Different precursors are tested as input for a statistical model to forecast high temperature events. Using multiple skill metrics for verification, we show that daily high temperature events have no predictive skill at long leads. By systematically testing the influence of temporal and spatial aggregation, we find that noise in the target time series is an important bottleneck for predicting extreme events on S2S time scales. We show that skill can be increased by a combination of 1) aggregating spatially and/or temporally, 2) lowering the threshold of the target events to increase the base rate, or 3) adding additional variables containing predictive information (soil moisture). Exploiting these skill-enhancing factors, we obtain forecast skill for moderate heat waves (i.e., 2 or more hot days closely clustered together in time) with up to 50 days of lead time.

Open access
Emiel van der Plas, Maurice Schmeits, Nicolien Hooijman, and Kees Kok

Abstract

Verification of localized events such as precipitation has become even more challenging with the advent of high-resolution mesoscale numerical weather prediction (NWP). The realism of a forecast suggests that it should compare well against precipitation radar imagery with similar resolution, both spatially and temporally. Spatial verification methods solve some of the representativity issues that point verification gives rise to. In this paper, a verification strategy based on model output statistics (MOS) is applied that aims to address both double-penalty and resolution effects that are inherent to comparisons of NWP models with different resolutions. Using predictors based on spatial precipitation patterns around a set of stations, an extended logistic regression (ELR) equation is deduced, leading to a probability forecast distribution of precipitation for each NWP model, analysis, and lead time. The ELR equations are derived for predictands based on areal-calibrated radar precipitation and SYNOP observations. The aim is to extract maximum information from a series of precipitation forecasts, like a trained forecaster would. The method is applied to the nonhydrostatic model Harmonie-AROME (2.5-km resolution), HIRLAM (11-km resolution), and the ECMWF model (16-km resolution), overall yielding similar Brier skill scores for the three postprocessed models, but somewhat larger differences for individual lead times. In addition, the fractions skill score is computed using the three deterministic forecasts, showing slightly higher skill for the Harmonie-AROME model. In other words, despite the realism of Harmonie-AROME precipitation forecasts, they only perform similarly or somewhat better than precipitation forecasts from the two lower-resolution models, at least in the Netherlands.

Full access
Simon Veldkamp, Kirien Whan, Sjoerd Dirksen, and Maurice Schmeits

Abstract

Current statistical postprocessing methods for probabilistic weather forecasting are not capable of using full spatial patterns from the numerical weather prediction (NWP) model. In this paper, we incorporate spatial wind speed information by using convolutional neural networks (CNNs) and obtain probabilistic wind speed forecasts in the Netherlands for 48 h ahead, based on KNMI’s deterministic HARMONIE-AROME NWP model. The probabilistic forecasts from the CNNs are shown to have higher Brier skill scores for medium to higher wind speeds, as well as a better continuous ranked probability score (CRPS) and logarithmic score, than the forecasts from fully connected neural networks and quantile regression forests. As a secondary result, we have compared the CNNs using three different density estimation methods [quantized softmax (QS), kernel mixture networks, and fitting a truncated normal distribution], and found the probabilistic forecasts based on the QS method to be best.

Restricted access