Since it is difficult to count on detailed information from ungauged basins so it can be used in a distributed model, a lumped model to forecast flood events have been identified and analyzed to be used in this type of basins. The Rainfall Threshold Curves method have been chosen for the present project due to its simplicity and low quantity demand of basin parameters. With the latter method, the threshold curves were calculated for 39 Italian sub-basins and later compared with the curves obtained by the FEST-WB distributed model to identify how different the results from both models were and therefore have an idea of the accuracy of the lumped model. Later, a comparison of the curves obtained from both models with 40 real events was made to understand how well the lumped model was at identifying heavy rainfall events in comparison with the FEST-WB model. The results were that from the 34 observed events which generated a flood, 21 and 19 were correctly identified by the lumped model and the distributed model respectively. Afterwards, a sensitivity analysis was made to identify which parameters were most important for the calculation of the threshold curves with the lumped model to subsequently use it in a real ungauged basin. The results were that the most important parameters were the CN and length of the stream with the maximum order. Finally, considering the information delivered by the previous analysis, the lumped model was applied for the calculation of the threshold curves on Bogotá River Basin which correctly identified more than half of the analyzed events.
Dato che è difficile avere informazione dettagliata da bacini non strumentati e quindi utilizzare modelli distribuiti, un modello concentrato è stato identificato ed analizzato pero la previsione di esondazione in questi tipi di bacini. Il metodo delle soglie pluviometriche come modello concentrato è stato scelto per la sua semplicità e l’uso di pochi parametri del bacino. Con questo metodo le curve di 39 bacini italiani sono state calcolate e poi confrontate con delle curve calcolate col modello distribuito FEST-WB per capire quanto diversi sono i risultati dei due modelli e finalmente per capire quanto affidabile è il modello concentrato. Poi, un confronto tra le curve dei due modelli e 40 eventi di piena è stato fatto per capire come si comporta il modello concentrato nel momento di identificare gli eventi. Il resultato è stato che per 34 eventi osservati che hanno generato una esondazione, 21 e 19 sono stati correttamente identificati dal modello concentrato e il modello distribuito rispettivamente. Dopo un’analisi di sensibilità è stato fatto per capire quali sono i parametri più importanti del modello concentrato per poi essere utilizzato in un bacino non-strumentato reale. Il risultato è stato che i parametri più importanti sono il CN e la lunghezza del ramo di ordine massimo. Finalmente, considerando le informazioni risultanti dalle analisi fatte, il modello concentrato è stato utilizzato per il calcolo delle soglie pluviometriche del Bacino del Fiume Bogotá. Le soglie risultanti sono state in grado di identificare correttamente metà degli eventi analizzati per il bacino colombiano.
Analysis of rainfall threshold curves for flood forecasting in ungauged basins and its application on a real case : Bogot river basin
LOPEZ TAMAYO, LUISA XIMENA
2016/2017
Abstract
Since it is difficult to count on detailed information from ungauged basins so it can be used in a distributed model, a lumped model to forecast flood events have been identified and analyzed to be used in this type of basins. The Rainfall Threshold Curves method have been chosen for the present project due to its simplicity and low quantity demand of basin parameters. With the latter method, the threshold curves were calculated for 39 Italian sub-basins and later compared with the curves obtained by the FEST-WB distributed model to identify how different the results from both models were and therefore have an idea of the accuracy of the lumped model. Later, a comparison of the curves obtained from both models with 40 real events was made to understand how well the lumped model was at identifying heavy rainfall events in comparison with the FEST-WB model. The results were that from the 34 observed events which generated a flood, 21 and 19 were correctly identified by the lumped model and the distributed model respectively. Afterwards, a sensitivity analysis was made to identify which parameters were most important for the calculation of the threshold curves with the lumped model to subsequently use it in a real ungauged basin. The results were that the most important parameters were the CN and length of the stream with the maximum order. Finally, considering the information delivered by the previous analysis, the lumped model was applied for the calculation of the threshold curves on Bogotá River Basin which correctly identified more than half of the analyzed events.File | Dimensione | Formato | |
---|---|---|---|
2017_04_Lopez_Tamayo.pdf
accessibile in internet per tutti
Descrizione: Tesi completa
Dimensione
6.87 MB
Formato
Adobe PDF
|
6.87 MB | Adobe PDF | Visualizza/Apri |
I documenti in POLITesi sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.
https://hdl.handle.net/10589/133670