Performance of GEM-LAM dichotomous forecast for selected weather phenomena

Małgorzata Zdunek , Maria Kłeczek

Abstract

In this study the performance of GEM-LAM numerical weather forecast focused on phenomena relevant to planning on-farm activities (i.e. frost and precipitation), is presented. Values from forecast were compared with observations gathered at 15 meteorological stations from Poland, for 1-year period. Based on data collected in contingency table, six verification scores were calculated. The results show that considerable bias exists – the model forecasts frost occurrence too rarely while precipitation events too frequently. However nearly half of frost cases and 2/3 of precipitation incidents were correctly predicted. As for success ratio SR score the frost forecast was more frequently correct (75% or 64%, depending on frost type) than forecast of precipitation (47%). The skill of negative forecast is high, especially for precipitation, where less than 5% of forecasts were erroneous. Analysis of verification scores calculated separately for each station shows, that regarding the forecast of frost, substantial differences in performance between particular locations exist. The worst results were obtained for stations located near the seaside which indicates that in the analyzed model the impact of water reservoir on frost formation is not correctly taken into account (at horizontal grid resolution of 5 km).
Author Małgorzata Zdunek KOKŚ
Małgorzata Zdunek,,
- Chair of Environmental Protection
, Maria Kłeczek
Maria Kłeczek,,
-
Journal seriesPrzegląd Naukowy Inżynieria i Kształtowanie Środowiska, ISSN 1732-9353
Issue year2016
Vol25
No4(74)
Pages483-496
Publication size in sheets0.65
Keywords in Englishdichotomous forecast, frost, precipitation, verification scores, GEM meteorological model
URL http://iks.pn.sggw.pl/PN74/A10/art10.pdf
Languageen angielski
Score (nominal)10
ScoreMinisterial score = 10.0, 28-11-2017, ArticleFromJournal
Ministerial score (2013-2016) = 10.0, 28-11-2017, ArticleFromJournal
Citation count*0
Cite
Share Share

Get link to the record
msginfo.png


* presented citation count is obtained through Internet information analysis and it is close to the number calculated by the Publish or Perish system.
Back