scholarly journals Value of uncertain streamflow observations for hydrological modelling

2018 ◽  
Vol 22 (10) ◽  
pp. 5243-5257 ◽  
Author(s):  
Simon Etter ◽  
Barbara Strobl ◽  
Jan Seibert ◽  
H. J. Ilja van Meerveld
Keyword(s):  
Time Series ◽  
Model Calibration ◽  
Model Performance ◽  
Error Distribution ◽  
Random Errors ◽  
Ungauged Catchments ◽  
Error Distributions ◽  
Data Points ◽  
Log Normal ◽  
Synthetic Time Series

Abstract. Previous studies have shown that hydrological models can be parameterised using a limited number of streamflow measurements. Citizen science projects can collect such data for otherwise ungauged catchments but an important question is whether these observations are informative given that these streamflow estimates will be uncertain. We assess the value of inaccurate streamflow estimates for calibration of a simple bucket-type runoff model for six Swiss catchments. We pretended that only a few observations were available and that these were affected by different levels of inaccuracy. The level of inaccuracy was based on a log-normal error distribution that was fitted to streamflow estimates of 136 citizens for medium-sized streams. Two additional levels of inaccuracy, for which the standard deviation of the error distribution was divided by 2 and 4, were used as well. Based on these error distributions, random errors were added to the measured hourly streamflow data. New time series with different temporal resolutions were created from these synthetic streamflow time series. These included scenarios with one observation each week or month, as well as scenarios that are more realistic for crowdsourced data that generally have an irregular distribution of data points throughout the year, or focus on a particular season. The model was then calibrated for the six catchments using the synthetic time series for a dry, an average and a wet year. The performance of the calibrated models was evaluated based on the measured hourly streamflow time series. The results indicate that streamflow estimates from untrained citizens are not informative for model calibration. However, if the errors can be reduced, the estimates are informative and useful for model calibration. As expected, the model performance increased when the number of observations used for calibration increased. The model performance was also better when the observations were more evenly distributed throughout the year. This study indicates that uncertain streamflow estimates can be useful for model calibration but that the estimates by citizen scientists need to be improved by training or more advanced data filtering before they are useful for model calibration.

scholarly journals Value of uncertain streamflow observations for hydrological modelling

2018 ◽  
Author(s):  
Simon Etter ◽  
Barbara Strobl ◽  
Jan Seibert ◽  
Ilja van Meerveld
Keyword(s):  
Time Series ◽  
Model Calibration ◽  
Model Performance ◽  
Error Distribution ◽  
Random Errors ◽  
Model Parameterization ◽  
Error Distributions ◽  
Log Normal ◽  
Synthetic Time Series ◽  
Runoff Model

Abstract. Previous studies have shown that a hydrological model can be parameterized using on a limited number of streamflow measurements for otherwise ungauged basins. Citizen science projects can collect such data but an important question is whether these observations are informative given that these streamflow estimates will be uncertain. We address the value of inaccurate streamflow estimates for calibration of a simple bucket-type runoff model for six Swiss catchments. We pretended that only a few observations were available and that these were affected by different levels of inaccuracy. The initial inaccuracy level was based on a log-normal error distribution that was fitted to streamflow estimates of 136 citizens for medium-sized streams. Two additional levels of inaccuracy, for which the standard deviation of the error-distribution was divided by two and four, were used as well. Based on these error distributions, random errors were added to the measured hourly streamflow data. New time series with different temporal resolutions were created from these synthetic time series. These included scenarios with one observation each week or month and scenarios that are more realistic for crowdsourced datasets with irregular distributions throughout the year or a focus on spring or summer. The model was then calibrated for the six catchments using the synthetic time series for a dry, an average and a wet year. The performance of the calibrated models was evaluated based on the measured hourly streamflow time series. The results indicate that streamflow estimates from untrained citizens are not informative for model calibration. However, if the errors can be reduced, the estimates are informative and useful for model parameterization. As expected, the model performance increased when the number of observations used for calibration increased. The model performance was also better when the observations were more evenly distributed throughout the year. This study indicates that uncertain streamflow estimates can be useful for model calibration but that the estimates by citizen scientists need to be improved by training or more advanced data filtering before they are useful for model calibration.

Statistical Properties of the Unsymmetrical P-Norm Distribution

2010 ◽  
Vol 439-440 ◽  
pp. 1153-1158
Author(s):  
Pan Xiong ◽  
Shuan Li Yuan ◽  
Shao Jie Cheng
Keyword(s):  
Distribution Function ◽  
Error Distribution ◽  
Statistical Properties ◽  
Test Method ◽  
Observation Error ◽  
Random Errors ◽  
True Error ◽  
Practical Situation ◽  
Error Distributions ◽  
The Common

The distribution of observation errors is determined according to their magnitudes by using the distribution collocation test method or figure method taking into account the result, sample total, the interval density etc. It is therefore difficult to get the specific type of error distribution of observations by conventional methods. In analyzing the actual situation of the observation error distribution using their statistical properties, this paper proposes the use of unsymmetrical distribution to express the true distribution of the observation errors. The P-norm distribution is a generalized form of a group of error distributions, and from the statistical properties of random errors we can arrive at an unsymmetrical P-norm distribution according to the practical situation of the occurrence of random errors. The common P-norm distribution is the specific case of this distribution. This paper deduces the density function equation of the unsymmetrical P-norm distribution, obtained the statistical properties of the distribution function and the evaluation of precision index. By choosing appropriate value for p, we can get closer to the distribution function of the true error distribution.

scholarly journals Assessment of the indirect calibration of a rainfall-runoff model for ungauged catchments in Flanders

2013 ◽  
Vol 17 (5) ◽  
pp. 2001-2016 ◽  
Author(s):  
N. De Vleeschouwer ◽  
V. R. N. Pauwels
Keyword(s):  
Time Series ◽  
Model Performance ◽  
Distributed Model ◽  
Calibration Data ◽  
Rainfall Runoff ◽  
Discharge Time ◽  
Ungauged Catchments ◽  
First Case ◽  
Spectral Calibration ◽  
The Time Domain

Abstract. In this paper the potential of discharge-based indirect calibration of the probability-distributed model (PDM), a lumped rainfall-runoff (RR) model, is examined for six selected catchments in Flanders. The concept of indirect calibration indicates that one has to estimate the calibration data because the catchment is ungauged or scarcely gauged. A first case in which indirect calibration is applied is that of spatial gauging divergence: because no observed discharge records are available at the outlet of the ungauged catchment, the calibration is carried out based on a rescaled discharge time series of a very similar donor catchment. Both a calibration in the time domain and the frequency domain (also known as spectral domain) are carried out. Furthermore, the case of temporal gauging divergence is considered: limited (e.g. historical or very recent) discharge records are available at the outlet of the scarcely gauged catchment. Additionally, no time overlap exists between the forcing and discharge records. Therefore, only an indirect spectral calibration can be performed in this case. To conclude also the combination case of spatio-temporal gauging divergence is considered. In this last case only limited discharge records are available at the outlet of a donor catchment. Again the forcing and discharge records are not concomitant, which only makes feasible an indirect spectral calibration. For most catchments the modelled discharge time series is found to be acceptable in the considered cases. In the case of spatial gauging divergence, indirect temporal calibration results in a better model performance than indirect spectral calibration. Furthermore, indirect spectral calibration in the case of temporal gauging divergence leads to a better model performance than indirect spectral calibration in the case of spatial gauging divergence. Finally, the combination of spatial and temporal gauging divergence does not lead to a notably worse model performance compared to the case of spatial gauging divergence.

scholarly journals Information content of stream level class data for hydrological model calibration

2017 ◽  
Vol 21 (9) ◽  
pp. 4895-4905 ◽  
Author(s):  
H. J. Ilja van Meerveld ◽  
Marc J. P. Vis ◽  
Jan Seibert
Keyword(s):  
Time Series ◽  
Citizen Science ◽  
Model Calibration ◽  
Hydrological Model ◽  
Model Performance ◽  
Added Value ◽  
Distributed Data ◽  
Vertical Resolution ◽  
Level Data ◽  
Spatially Distributed

Abstract. Citizen science can provide spatially distributed data over large areas, including hydrological data. Stream levels are easier to measure than streamflow and are likely also observed more easily by citizen scientists than streamflow. However, the challenge with crowd based stream level data is that observations are taken at irregular time intervals and with a limited vertical resolution. The latter is especially the case at sites where no staff gauge is available and relative stream levels are observed based on (in)visible features in the stream, such as rocks. In order to assess the potential value of crowd based stream level observations for model calibration, we pretended that stream level observations were available at a limited vertical resolution by transferring streamflow data to stream level classes. A bucket-type hydrological model was calibrated with these hypothetical stream level class data and subsequently evaluated on the observed streamflow records. Our results indicate that stream level data can result in good streamflow simulations, even with a reduced vertical resolution of the observations. Time series of only two stream level classes, e.g. above or below a rock in the stream, were already informative, especially when the class boundary was chosen towards the highest stream levels. There was some added value in using up to five stream level classes, but there was hardly any improvement in model performance when using more level classes. These results are encouraging for citizen science projects and provide a basis for designing observation systems that collect data that are as informative as possible for deriving model based streamflow time series for previously ungauged basins.

scholarly journals Streamflow characteristics from modeled runoff time series – importance of calibration criteria selection

2017 ◽  
Vol 21 (11) ◽  
pp. 5443-5457 ◽  
Author(s):  
Sandra Pool ◽  
Marc J. P. Vis ◽  
Rodney R. Knight ◽  
Jan Seibert
Keyword(s):  
Time Series ◽  
Model Calibration ◽  
Performance Metrics ◽  
Low Flow ◽  
Flow Conditions ◽  
Ungauged Catchments ◽  
Calibration Process ◽  
Model Efficiency ◽  
Runoff Model ◽  
Calibration Approach

Abstract. Ecologically relevant streamflow characteristics (SFCs) of ungauged catchments are often estimated from simulated runoff of hydrologic models that were originally calibrated on gauged catchments. However, SFC estimates of the gauged donor catchments and subsequently the ungauged catchments can be substantially uncertain when models are calibrated using traditional approaches based on optimization of statistical performance metrics (e.g., Nash–Sutcliffe model efficiency). An improved calibration strategy for gauged catchments is therefore crucial to help reduce the uncertainties of estimated SFCs for ungauged catchments. The aim of this study was to improve SFC estimates from modeled runoff time series in gauged catchments by explicitly including one or several SFCs in the calibration process. Different types of objective functions were defined consisting of the Nash–Sutcliffe model efficiency, single SFCs, or combinations thereof. We calibrated a bucket-type runoff model (HBV – Hydrologiska Byråns Vattenavdelning – model) for 25 catchments in the Tennessee River basin and evaluated the proposed calibration approach on 13 ecologically relevant SFCs representing major flow regime components and different flow conditions. While the model generally tended to underestimate the tested SFCs related to mean and high-flow conditions, SFCs related to low flow were generally overestimated. The highest estimation accuracies were achieved by a SFC-specific model calibration. Estimates of SFCs not included in the calibration process were of similar quality when comparing a multi-SFC calibration approach to a traditional model efficiency calibration. For practical applications, this implies that SFCs should preferably be estimated from targeted runoff model calibration, and modeled estimates need to be carefully interpreted.

scholarly journals Objective functions used as performance metrics for hydrological models: state-of-the-art and critical analysis

RBRH ◽  
2020 ◽  
Vol 25 ◽  
Author(s):  
Paloma Mara de Lima Ferreira ◽  
Adriano Rolim da Paz ◽  
Juan Martín Bravo
Keyword(s):  
Time Series ◽  
Model Calibration ◽  
Hydrological Model ◽  
Performance Metrics ◽  
Low Flow ◽  
Hydrological Models ◽  
Objective Functions ◽  
Flow Rates ◽  
Daily Streamflow ◽  
Synthetic Time Series

ABSTRACT Hydrological models (HMs) can be applied for different purposes, and a key step is model calibration using objective functions (OF) to quantify the agreement between observed and calculated discharges. Fully understanding the OF is important to properly take advantage of model calibration and interpret the results. This study evaluates 36 OF proposed in the literature, considering two watersheds of different hydrological regimes. Daily simulated streamflow time-series, using a distributed hydrological model (MGB-IPH), and ten daily streamflow synthetic time-series, generated from the observed and calculated streamflows, were used in the analysis of each watershed. These synthetic data were used to evaluate how does each metric evaluate hypothetical cases that present isolated very well known error behaviors. Despite of all NSE-derived (Nash-Sutcliffe efficiency) metrics that use the square of the residuals in their formulation have shown higher sensitivity to errors in high flows, the ones that use daily and monthly averages of flow rates in absolute terms were more stringent than the others to assess HMs performance. Low flow errors were better evaluated by metrics that use the flow logarithm. The constant presence of zero flow rates deteriorate them significantly, with the exception of the metrics TRMSE (Transformed root mean square error) did not demonstrate this problem. An observed limitation of the formulations of some metrics was that the errors of overestimation or underestimation are compensated. Our results reassert that each metric should be interpreted specifically thinking about the aspects it has been proposed for, and simultaneously taking into account a set of metrics would lead to a broader evaluation of HM ability (e.g. multiobjective model evaluation). We recommend that the use of synthetic time series as those proposed in this work could be useful as an auxiliary step towards better understanding the evaluation of a calibrated hydrological model for each study case, taking into account model capabilities and observed hydrologic regime characteristics.

scholarly journals Information content of stream level class data for hydrological model calibration

2017 ◽  
Author(s):  
Ilja van Meerveld ◽  
Marc Vis ◽  
Jan Seibert
Keyword(s):  
Time Series ◽  
Citizen Science ◽  
Model Calibration ◽  
Hydrological Model ◽  
Model Performance ◽  
Added Value ◽  
Data Sets ◽  
Distributed Data ◽  
Vertical Resolution ◽  
Level Data

Abstract. Citizen science can provide spatially distributed data over large areas, including hydrological data. Stream levels are easier to measure than streamflow and can be observed more easily by citizen scientists. However, the challenge with crowd-based stream level data is that observations are taken at irregular time intervals and with a limited vertical resolution. The latter is especially the case at sites where no staff gauge is available and relative stream levels are observed based on (in)visible features in the stream, such as rocks. In order to assess the potential value of crowd-based stream level observations for model calibration, we pretended that stream level observations were available at a limited vertical resolution by transferring streamflow data into stream level classes. A bucket-type hydrological model was calibrated with these hypothetical data sets and subsequently evaluated on the observed streamflow records. Our results indicate that stream level data can result in good streamflow simulations, even with a reduced vertical resolution of the observations. Time series of only two stream level classes, e.g. above or below a rock in the stream, were already informative, especially when the class boundary was chosen towards the highest stream levels. There was some added value in using up to five stream level classes but there was hardly any improvement in model performance when using more level classes. These results are encouraging for citizen science projects and provide a basis for designing observation systems that collect data that are as informative as possible for deriving model-based streamflow time series for previously ungauged basins.

scholarly journals Streamflow characteristics from modelled runoff time series – importance of calibration criteria selection

2016 ◽  
Author(s):  
Sandra Pool ◽  
Marc J. P. Vis ◽  
Rodney R. Knight ◽  
Jan Seibert
Keyword(s):  
Model Calibration ◽  
Performance Metrics ◽  
Model Performance ◽  
Specific Model ◽  
Low Flow ◽  
Flow Conditions ◽  
Ungauged Catchments ◽  
Practical Applications ◽  
Runoff Model ◽  
Calibration Approach

Abstract. Ecologically relevant streamflow characteristics (SFCs) of ungauged catchments are often estimated from simulated runoff of hydrologic models. Estimated SFCs can be substantially uncertain when models are calibrated using traditional approaches based on minimization or maximization of statistical performance metrics (e.g. Nash–Sutcliffe efficiency). To evaluate model performance, we tested how well SFCs are simulated when the model objective function was calibrated using one or more SFCs. We calibrated a bucket-type runoff model for 25 catchments in the Tennessee River basin and evaluated the proposed calibration approach on 13 selected SFCs representing major flow regime components and different flow conditions. While the model tends to underestimate SFCs related to mean and high-flow conditions, SFCs related to low flow are overestimated. The highest estimation accuracies were achieved by a SFC-specific model calibration. Estimates of SFCs not included in the calibration process were of similar quality when comparing a multi-SFC calibration approach to a traditional Nash–Sutcliffe efficiency calibration. For practical applications, this implies that SFCs should preferably be estimated from targeted runoff model calibration and modelled estimates need to be carefully interpreted.

scholarly journals Assessment of the indirect calibration of a rainfall-runoff model for ungauged catchments in Flanders

2013 ◽  
Vol 10 (1) ◽  
pp. 103-144
Author(s):  
N. De Vleeschouwer ◽  
V. R. N. Pauwels
Keyword(s):  
Time Series ◽  
Model Performance ◽  
Distributed Model ◽  
Calibration Data ◽  
Rainfall Runoff ◽  
Discharge Time ◽  
Ungauged Catchments ◽  
Ungauged Catchment ◽  
First Case ◽  
Spectral Calibration

Abstract. In this paper the potential of discharge-based indirect calibration of the Probability Distributed Model (PDM), a lumped rainfall-runoff (RR) model, is examined for six selected catchments in Flanders. The concept of indirect calibration indicates that one has to estimate the calibration data because the catchment is ungauged. A first case in which indirect calibration is applied is that of spatial gauging divergence: Because no observed discharge records are available at the outlet of the ungauged catchment, the calibration is carried out based on a rescaled discharge time series of a very similar donor catchment. Both a calibration in the time domain and the frequency domain (a.k.a. spectral domain) are carried out. Furterhermore, the case of temporal gauging divergence is considered: Limited (e.g. historical or very recent) discharge records are available at the outlet of the ungauged catchment. Additionally, no time overlap exists between the forcing and discharge records. Therefore, only an indirect spectral calibration can be performed in this case. To conclude also the combination case of spatio-temporal gauging divergence is considered. In this last case only limited discharge records are available at the outlet of a donor catchment. Again the forcing and discharge records are not contemporaneous which only makes feasible an indirect spectral calibration. The modelled discharge time series are found to be acceptable in all three considered cases. In the case of spatial gauging divergence, indirect temporal calibration results in a slightly better model performance than indirect spectral calibration. Furthermore, indirect spectral calibration in the case of temporal gauging divergence leads to a better model performance than indirect spectral calibration in the case of spatial gauging divergence. Finally, the combination of spatial and temporal gauging divergence does not necessarily lead to a worse model performance compared to the separate cases of spatial and temporal gauging divergence.

scholarly journals A new approach to model the variability of karstic recharge

2012 ◽  
Vol 9 (2) ◽  
pp. 2443-2473
Author(s):  
A. Hartmann ◽  
J. Lange ◽  
M. Weiler ◽  
Y. Arbel ◽  
N. Greenbaum
Keyword(s):  
Time Series ◽  
Spatial Variability ◽  
Model Calibration ◽  
Model Performance ◽  
Spatial And Temporal Variability ◽  
Hydrological Models ◽  
New Approach ◽  
Perceptual Model ◽  
Spatially Distributed ◽  
Karst Systems

Abstract. In karst systems, surface near dissolution carbonate rock results in a high spatial and temporal variability of groundwater recharge. To adequately represent the dominating recharge processes in hydrological models is still a challenge, especially in data scare regions. In this study, we developed a recharge model that is based on a perceptual model of the epikarst. It represents epikarst heterogeneity as a set of system property distributions to produce not only a single recharge time series, but a variety of time series representing the spatial recharge variability. We tested the new model with a unique set of spatially distributed flow and tracer observations in a karstic cave at Mt. Carmel, Israel. We transformed the spatial variability into statistical variables and apply an iterative calibration strategy in which more and more data was added to the calibration. Thereby, we could show that the model is only able to produce realistic results when the information about the spatial variability of the observations was included into the model calibration. We could also show that tracer information improves the model performance if data about the variability is not included.

Sign in / Sign up

Export Citation Format

Share Document