On the number of bins in a rank histogram

Heinrich, Claudio

doi:10.1002/qj.3932

Cited by 8 publications

(9 citation statements)

References 17 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…The PIT (probability integral transform) for a day is the fraction of ensemble members (here area‐mean precipitation of E‐OBS‐Ens) smaller than the reference (the deterministic estimate of APGD), and the histogram summarizes PITs from all wet days in 1979–2008. A statistically reliable ensemble yields uniformly distributed PITs (Hamill, 2001; Heinrich, 2020). Here, prior to calculating the PITs, we have adjusted the monthly long‐term mean values of E‐OBS‐Ens to be the same as those of APGD (multiplicative bias correction).…”

Section: Resultsmentioning

confidence: 99%

Evaluation of daily precipitation analyses in E‐OBS (v19.0e) and ERA5 by comparison to regional high‐resolution datasets in European regions

Bandhauer

Isotta

Lakatos

et al. 2021

Intl Journal of Climatology

106

View full text Add to dashboard Cite

Gridded analyses of observed precipitation are an important data resource for environmental modelling, climate model evaluation and climate monitoring.In Europe, datasets that resolve the rich mesoscale variations widely exist for the national territories, but similar datasets covering the entire continent are more recent. Here, we evaluate daily precipitation in two newly available pan-European datasets: E-OBS (v19.0e), a statistical analysis from rain-gauge data, and ERA5, the new global reanalysis from ECMWF. Special interest is on how the refinements of grid spacing, the methodological upgrades and the quantification of uncertainty (ensemble), bear on capabilities at the mesoscale.The evaluation is conducted in three subregions, the Alps, the Carpathians and Fennoscandia, and involves as reference high-quality regional datasets derived from dense rain-gauge data. The study suggests that E-OBS and ERA5 agree qualitatively well with the reference datasets. Major mesoscale patterns in the climatology (mean, wet-day frequency, 95% quantile) are reproduced.The improvement over earlier versions of the datasets is evident. ERA5 was found to overestimate mean precipitation in all regions, related to too many wet days. The accuracy of E-OBS was found to depend on station density, with spatial and temporal variations clearly less accurate in data sparse regions. In comparison, E-OBS turned out to be superior to ERA5 in regions with dense data, but the two datasets are on a par in regions with sparse data, and partly ERA5 has advantages. For both datasets we find that the spatial resolution is coarser than the grid spacing, with overly smooth fields and an underestimation of high quantiles. Also, both datasets were found to be clearly overconfident in their uncertainty characterization (too small ensemble spread).Overall, the two datasets advance the characterization of precipitation on a

show abstract

Section: Resultsmentioning

confidence: 99%

Evaluation of daily precipitation analyses in E‐OBS (v19.0e) and ERA5 by comparison to regional high‐resolution datasets in European regions

Bandhauer

Isotta

Lakatos

et al. 2021

Intl Journal of Climatology

106

View full text Add to dashboard Cite

show abstract

“…The calibration assessment is performed by calculating the rank of each seNorge observation within the corresponding forecast ensemble, and building rank histograms (Hamill, 2001). To facilitate comparison across forecast cases with varying ensemble size, we standardize the observed ranks to take values between 0 and 1 by subtracting 1 and then dividing by the ensemble size, with ties resolved at random (Heinrich, 2021).…”

Section: Calibration Assessmentmentioning

confidence: 99%

Probabilistic prediction of the time to hard freeze using seasonal weather forecasts and survival time methods

Roksvåg¹,

Lenkoski²,

Scheuerer³

et al. 2022

Preprint

View full text Add to dashboard Cite

Agricultural food production and natural ecological systems depend on a range of seasonal climate indicators that describe seasonal patterns in climatological conditions. This paper proposes a probabilistic forecasting framework for predicting the end of the freeze-free season, or the time to a mean daily near-surface air temperature below 0 • C (here referred to as hard freeze). The forecasting framework is based on the multi-model seasonal forecast ensemble provided by the Copernicus Climate Data Store and uses techniques from survival analysis for time-to-event data. The original mean daily temperature forecasts are statistically post-processed with a mean and variance correction of each model system before the time-to-event forecast is constructed. In a case study for a region in Fennoscandia covering Norway for the period 1993-2020, the proposed forecasts are found to outperform a climatology forecast from an observation-based data product at locations where the average predicted time to hard freeze is less than 40 days after the initialization date of the forecast on October 1.

show abstract

“…To facilitate comparison across forecast cases with varying ensemble size, we standardize the observed ranks to take values between 0 and 1 by subtracting 1 and then dividing by the ensemble size (Heinrich, 2021). For a calibrated forecast, the observed standardized ranks should be uniformly distributed on [0, 1].…”

Section: Calibration Assessmentmentioning

confidence: 99%

“…Hence, it is pretty common that an observation takes the exact same value as one or several of the ensemble members. If this is not taken into account, one gets ties in the ranking that can lead to skewed probability integral transform histograms even if the ranks are perfectly uniformly distributed (Heinrich, 2021). To solve the problem, we add normal‐distributed noise to the data such that ties are resolved at random.…”

Section: Forecast Evaluationmentioning

confidence: 99%

Probabilistic prediction of the time to hard freeze using seasonal weather forecasts and survival time methods

Roksvåg

Lenkoski

Scheuerer

et al. 2022

Quart J Royal Meteoro Soc

View full text Add to dashboard Cite

Agricultural food production and natural ecological systems depend on a range of seasonal climate indicators that describe seasonal patterns in climatological conditions. This article proposes a probabilistic forecasting framework for predicting the end of the freeze‐free season, or the time to a mean daily near‐surface air temperature below 0°C (referred to here as hard freeze). The forecasting framework is based on the multimodel seasonal forecast ensemble provided by the Copernicus Climate Data Store and uses techniques from survival analysis for time‐to‐event data. The original mean daily temperature forecasts are statistically postprocessed and downscaled with a mean and variance correction of each model system before the time‐to‐event forecast is constructed. In a case study for a region in Fennoscandia covering Norway for the period 1993–2020, the proposed forecasts are found to outperform a climatology forecast from an observation‐based data product at locations where the average predicted time to hard freeze is less than 40 days after the initialization date of the forecast on October 1.

show abstract

On the number of bins in a rank histogram

Cited by 8 publications

References 17 publications

Evaluation of daily precipitation analyses in E‐OBS (v19.0e) and ERA5 by comparison to regional high‐resolution datasets in European regions

Evaluation of daily precipitation analyses in E‐OBS (v19.0e) and ERA5 by comparison to regional high‐resolution datasets in European regions

Probabilistic prediction of the time to hard freeze using seasonal weather forecasts and survival time methods

Probabilistic prediction of the time to hard freeze using seasonal weather forecasts and survival time methods

Contact Info

Product

Resources

About