Evaluation Of Statistical Models For Forecast Errors from the HBV-Model
Three statistical models for the forecast errors for inflow to the Langvatn reservoir in Northern Norway have been constructed and tested according to how well the distribution and median values of the forecasts errors fit to the observations. For the first model observed and forecasted inflows were...
Main Authors: | , , , |
---|---|
Other Authors: | , |
Format: | Still Image |
Language: | English |
Published: |
2009
|
Subjects: | |
Online Access: | https://irsteadoc.irstea.fr/cemoa/PUB00027508 |
id |
ftcemoa:oai:irsteadoc.irstea.fr:PUB00027508 |
---|---|
record_format |
openpolar |
spelling |
ftcemoa:oai:irsteadoc.irstea.fr:PUB00027508 2023-05-15T17:43:36+02:00 Evaluation Of Statistical Models For Forecast Errors from the HBV-Model Engeland, K. Kolberg, S. Renard, B. Stensland, I. SINTEF TRONDHEIM NOR CEMAGREF LYON UR HHLY FRA 2009 application/pdf https://irsteadoc.irstea.fr/cemoa/PUB00027508 Anglais eng https://irsteadoc.irstea.fr/cemoa/PUB00027508 Date de dépôt: 2009-12-21 - Tous les documents et informations contenus dans la base CemOA Publications sont protégés en vertu du droit de propriété intellectuelle, en particulier par le droit d'auteur. La personne consultant la base CemOA Publications peut visualiser, reproduire, ou stocker des copies des publications, à condition que l'information soit seulement pour son usage personnel et non commercial. L'utilisation des travaux universitaires est soumise à autorisation préalable de leurs auteurs. Toute information relative au signalement d'une publication contenue dans CemOA Publications doit inclure la citation bibliographique usuelle : Nom du ou des auteurs, titre et source du document, date et URL de la notice (dc_identifier). 22895 MODELE STATISTIQUE PREVISION HBV FORECAST STATISTICAL MODEL Poster 2009 ftcemoa 2021-06-29T09:21:56Z Three statistical models for the forecast errors for inflow to the Langvatn reservoir in Northern Norway have been constructed and tested according to how well the distribution and median values of the forecasts errors fit to the observations. For the first model observed and forecasted inflows were transformed by the Box-Cox transformation before a first order autoregressive model was constructed for the forecast errors. The parameters were conditioned on climatic conditions. In the second model the Normal Quantile Transformation (NQT) was applied on observed and forecasted inflows before a similar first order autoregressive model was constructed for the forecast errors. For the last model positive and negative errors were modeled separately. The errors were first NQT-transformed before a model where the mean values were conditioned on climate, forecasted inflow and yesterday's error. To test the three models we applied three criterions: We wanted a) the median values to be close to the observed values; b) the forecast intervals to be narrow; c) the distribution to be correct. The results showed that it is difficult to obtain a correct model for the forecast errors, and that the main challenge is to account for the auto-correlation in the errors. Model 1 and 2 gave similar results, and the main drawback is that the distributions are not correct. Model 3 gave a distribution that fits better, but the median values do not fit well since the auto-correlation is not properly accounted for. If the 95% forecast interval is of interest, Model 2 is recommended. If the whole distribution is of interest, Model 3 is recommended. Still Image Northern Norway Irstea Publications et Bases documentaires (Irstea@doc/CemOA) Norway |
institution |
Open Polar |
collection |
Irstea Publications et Bases documentaires (Irstea@doc/CemOA) |
op_collection_id |
ftcemoa |
language |
English |
topic |
MODELE STATISTIQUE PREVISION HBV FORECAST STATISTICAL MODEL |
spellingShingle |
MODELE STATISTIQUE PREVISION HBV FORECAST STATISTICAL MODEL Engeland, K. Kolberg, S. Renard, B. Stensland, I. Evaluation Of Statistical Models For Forecast Errors from the HBV-Model |
topic_facet |
MODELE STATISTIQUE PREVISION HBV FORECAST STATISTICAL MODEL |
description |
Three statistical models for the forecast errors for inflow to the Langvatn reservoir in Northern Norway have been constructed and tested according to how well the distribution and median values of the forecasts errors fit to the observations. For the first model observed and forecasted inflows were transformed by the Box-Cox transformation before a first order autoregressive model was constructed for the forecast errors. The parameters were conditioned on climatic conditions. In the second model the Normal Quantile Transformation (NQT) was applied on observed and forecasted inflows before a similar first order autoregressive model was constructed for the forecast errors. For the last model positive and negative errors were modeled separately. The errors were first NQT-transformed before a model where the mean values were conditioned on climate, forecasted inflow and yesterday's error. To test the three models we applied three criterions: We wanted a) the median values to be close to the observed values; b) the forecast intervals to be narrow; c) the distribution to be correct. The results showed that it is difficult to obtain a correct model for the forecast errors, and that the main challenge is to account for the auto-correlation in the errors. Model 1 and 2 gave similar results, and the main drawback is that the distributions are not correct. Model 3 gave a distribution that fits better, but the median values do not fit well since the auto-correlation is not properly accounted for. If the 95% forecast interval is of interest, Model 2 is recommended. If the whole distribution is of interest, Model 3 is recommended. |
author2 |
SINTEF TRONDHEIM NOR CEMAGREF LYON UR HHLY FRA |
format |
Still Image |
author |
Engeland, K. Kolberg, S. Renard, B. Stensland, I. |
author_facet |
Engeland, K. Kolberg, S. Renard, B. Stensland, I. |
author_sort |
Engeland, K. |
title |
Evaluation Of Statistical Models For Forecast Errors from the HBV-Model |
title_short |
Evaluation Of Statistical Models For Forecast Errors from the HBV-Model |
title_full |
Evaluation Of Statistical Models For Forecast Errors from the HBV-Model |
title_fullStr |
Evaluation Of Statistical Models For Forecast Errors from the HBV-Model |
title_full_unstemmed |
Evaluation Of Statistical Models For Forecast Errors from the HBV-Model |
title_sort |
evaluation of statistical models for forecast errors from the hbv-model |
publishDate |
2009 |
url |
https://irsteadoc.irstea.fr/cemoa/PUB00027508 |
geographic |
Norway |
geographic_facet |
Norway |
genre |
Northern Norway |
genre_facet |
Northern Norway |
op_source |
22895 |
op_relation |
https://irsteadoc.irstea.fr/cemoa/PUB00027508 |
op_rights |
Date de dépôt: 2009-12-21 - Tous les documents et informations contenus dans la base CemOA Publications sont protégés en vertu du droit de propriété intellectuelle, en particulier par le droit d'auteur. La personne consultant la base CemOA Publications peut visualiser, reproduire, ou stocker des copies des publications, à condition que l'information soit seulement pour son usage personnel et non commercial. L'utilisation des travaux universitaires est soumise à autorisation préalable de leurs auteurs. Toute information relative au signalement d'une publication contenue dans CemOA Publications doit inclure la citation bibliographique usuelle : Nom du ou des auteurs, titre et source du document, date et URL de la notice (dc_identifier). |
_version_ |
1766145716036567040 |