Aplicación Práctica De Modelaje Hidrológico En Panamá
Páginas: 5 (1134 palabras)
Publicado: 13 de julio de 2011
MSC IN HYDROLOGY
University of Costa Rica
University of Oslo
GEO 4320
HYDROLOGICAL MODELLING
ASSIGMENT #1
HBV Model
Submitted by
BLANCA TAMARA SOLIS G.
1. Introduction
A model is a representation, or description of how works the nature in some specific process and help us to understand it behavoir.
Investigating on diferent sites on the internet and publications, I couldfind that there is a big number of hydrological models and depending on the hydrological model and what products do you need to reproduce, there are model with some specific input parameters and corresponding output.
At this report, we are calibrating manually the model and then verifying the model’s parameters by using the simple split-sample method; and finally, using Monte Carlo Options, we arestudying the sensitivity of these parameters.
HBV Model is a Conceptual Model, a model between theoretical and empirical. It take considerations on physical laws but in a simplified way. During the last 20 years the HBV model has become widely used for runoff simulations in Sweden. The model has been applied, sometimes in modified versions, in about 30 countries.
2. Data
The data used wasthe daily discharge, average daily discharge and temperature is taken from the catchment, Akesta at Kvarn in Central Sweden.
I decided to use this data because at that moment there was a problem on getting the data from the server in ETESA. Other reason was because I started to prepare the assigment when I was on Costa Rica. Now, this problem is solved and we can work on panamanian data for thefuture.
3. Procedure
• Calibrate the HBV model with Manual calibration method. It is necessary to start with a warming up period and, it most be calibrated for another period.
• Verify the model parameter values by using the simple split-sample method.
• Run Monte Carlo Options to study de sensitivity and uncertainty of the model parameters.
4. Outputs and their AnalysisCalibration
Period: from september 01, 1981 until september 01, 1982 (Figure 1)
and september 01, 1982 until september 01, 1983 (Figure 2)
Giving manually different values to the parameters, we found a better value for R^2. The end value of R^2 was 0.9019 after the manual calibration of parameters and the efficiency of the model was 0.8278. There are shown, on the table, the new values for allparameters.
End values
FC [mm] 700
LP [-] 0.6
BETA [-] 5
MAXBAS [d] 1
CET [1/C] 0.1
PERC [mm/d] 100
UZL [mm] 0
K0 [l/d] O
K1 [l/d] 0.1
K2 [l/d] 0.05
R2 0.9019
Table 1. Parameters obtained after Manual Calibration
Figure 1. Comparison between observed and simulated values.
The peaks on Q calculated responds to the peaks of the observed data. It doesn’t respond exactly but hasthe same tendency.
The analyzed period shown on the figure was from September 01/1981 to September 01/1982.
Figure 2. Comparison between observed and calculated Q.
On figure 2 can observe that the tendency of the observed and calculated is the same. There is a good reproduction of that values, specially, the peak values. The analyzed period shown on the figure 2 was from September01/1982 to September 01/1983.
Validation
Period: from august 31, 1987 until december 12, 1991 (Figure 3)
After finding good parameter values from one data period, we run the model for another independent period (without changing the parameter values).
Figure 3. Model Efficiency value and R value for the new period.
The R^2 decreases little but remains good value and the modelefficiency keeps good value too. 0.7737 and 0.8478 respectively.
Figure 3. The Q calculated responds to the peaks of the observed data for these period.
Figure 4. The Q calculated responds to the peaks of the observed data.
The chosen parameters were applied for both periods and was found that for the last period, the efficiency decreases a little but it remains with good value....
University of Costa Rica
University of Oslo
GEO 4320
HYDROLOGICAL MODELLING
ASSIGMENT #1
HBV Model
Submitted by
BLANCA TAMARA SOLIS G.
1. Introduction
A model is a representation, or description of how works the nature in some specific process and help us to understand it behavoir.
Investigating on diferent sites on the internet and publications, I couldfind that there is a big number of hydrological models and depending on the hydrological model and what products do you need to reproduce, there are model with some specific input parameters and corresponding output.
At this report, we are calibrating manually the model and then verifying the model’s parameters by using the simple split-sample method; and finally, using Monte Carlo Options, we arestudying the sensitivity of these parameters.
HBV Model is a Conceptual Model, a model between theoretical and empirical. It take considerations on physical laws but in a simplified way. During the last 20 years the HBV model has become widely used for runoff simulations in Sweden. The model has been applied, sometimes in modified versions, in about 30 countries.
2. Data
The data used wasthe daily discharge, average daily discharge and temperature is taken from the catchment, Akesta at Kvarn in Central Sweden.
I decided to use this data because at that moment there was a problem on getting the data from the server in ETESA. Other reason was because I started to prepare the assigment when I was on Costa Rica. Now, this problem is solved and we can work on panamanian data for thefuture.
3. Procedure
• Calibrate the HBV model with Manual calibration method. It is necessary to start with a warming up period and, it most be calibrated for another period.
• Verify the model parameter values by using the simple split-sample method.
• Run Monte Carlo Options to study de sensitivity and uncertainty of the model parameters.
4. Outputs and their AnalysisCalibration
Period: from september 01, 1981 until september 01, 1982 (Figure 1)
and september 01, 1982 until september 01, 1983 (Figure 2)
Giving manually different values to the parameters, we found a better value for R^2. The end value of R^2 was 0.9019 after the manual calibration of parameters and the efficiency of the model was 0.8278. There are shown, on the table, the new values for allparameters.
End values
FC [mm] 700
LP [-] 0.6
BETA [-] 5
MAXBAS [d] 1
CET [1/C] 0.1
PERC [mm/d] 100
UZL [mm] 0
K0 [l/d] O
K1 [l/d] 0.1
K2 [l/d] 0.05
R2 0.9019
Table 1. Parameters obtained after Manual Calibration
Figure 1. Comparison between observed and simulated values.
The peaks on Q calculated responds to the peaks of the observed data. It doesn’t respond exactly but hasthe same tendency.
The analyzed period shown on the figure was from September 01/1981 to September 01/1982.
Figure 2. Comparison between observed and calculated Q.
On figure 2 can observe that the tendency of the observed and calculated is the same. There is a good reproduction of that values, specially, the peak values. The analyzed period shown on the figure 2 was from September01/1982 to September 01/1983.
Validation
Period: from august 31, 1987 until december 12, 1991 (Figure 3)
After finding good parameter values from one data period, we run the model for another independent period (without changing the parameter values).
Figure 3. Model Efficiency value and R value for the new period.
The R^2 decreases little but remains good value and the modelefficiency keeps good value too. 0.7737 and 0.8478 respectively.
Figure 3. The Q calculated responds to the peaks of the observed data for these period.
Figure 4. The Q calculated responds to the peaks of the observed data.
The chosen parameters were applied for both periods and was found that for the last period, the efficiency decreases a little but it remains with good value....
Leer documento completo
Regístrate para leer el documento completo.