Applications of digital baseflow separation techniques for model validation, Wairarapa valley, New Zealand

<p>The representation of groundwater processes in hydrological models is crucial, as the connectivity between groundwater and surface water is significant. It is particularly important for regions such as the Wairarapa that experience high water stresses. Intensified agriculture has increased demand for irrigation, which can lead to depletion and degradation of reservoirs. This study compared observed streamflow records to TopNet-0 and TopNet-GW model outputs at points along the Mangatarere stream, a sub-catchment in the Wairarapa valley, New Zealand. Model performance was assessed using a suite of quantitative and qualitative comparisons. This analysis aimed to assess the similarities and differences between observed flow and the model outputs with respect to their model structures. Baseflow estimates from recursive digital filters were also compared at these sites to assess the groundwater representation of the models. The investigation can be considered representative of the wider Ruamahanga catchment, as the geology and hydrology in the region is relatively analogous. Flow infilling and baseflow separation was undertaken at 13 Wairarapa flow gauges to provide considerations to the model outputs. Options investigated for flow infilling included a straight infill or calculation of the flow difference at each point. Potential multipliers included a long-term or a monthly option. The difference infill, coupled with the long-term multiplier, was found to be the optimum method. Independent baseflow estimates included a Q90/Q50 flow duration curve index and indices generated from the Eckhardt and Bump & Rise recursive digital filters. The two digital filters produced similar statistics but were found to employ uncertain parameters that significantly affect outputs. TopNet-GW benefitted from up-to-date calibrations and as such produced generally excellent simulations in comparison to observed streamflow. With the addition of the deep groundwater conceptual reservoir in the structure of the model, simulated flow baseflow index estimates and graphical assessment of flow recession curves indicate TopNet-GW reproduces groundwater processes well despite potential over-representation of baseflow at the expense of high flow periods during peak flows. These findings highlight the importance of combining subsurface and surface flow dynamics to resolve water management issues and improve model performance at the catchment scale.</p>

Applications of digital baseflow separation techniques for model validation, Wairarapa valley, New Zealand

<p>The representation of groundwater processes in hydrological models is crucial, as the connectivity between groundwater and surface water is significant. It is particularly important for regions such as the Wairarapa that experience high water stresses. Intensified agriculture has increased demand for irrigation, which can lead to depletion and degradation of reservoirs. This study compared observed streamflow records to TopNet-0 and TopNet-GW model outputs at points along the Mangatarere stream, a sub-catchment in the Wairarapa valley, New Zealand. Model performance was assessed using a suite of quantitative and qualitative comparisons. This analysis aimed to assess the similarities and differences between observed flow and the model outputs with respect to their model structures. Baseflow estimates from recursive digital filters were also compared at these sites to assess the groundwater representation of the models. The investigation can be considered representative of the wider Ruamahanga catchment, as the geology and hydrology in the region is relatively analogous. Flow infilling and baseflow separation was undertaken at 13 Wairarapa flow gauges to provide considerations to the model outputs. Options investigated for flow infilling included a straight infill or calculation of the flow difference at each point. Potential multipliers included a long-term or a monthly option. The difference infill, coupled with the long-term multiplier, was found to be the optimum method. Independent baseflow estimates included a Q90/Q50 flow duration curve index and indices generated from the Eckhardt and Bump & Rise recursive digital filters. The two digital filters produced similar statistics but were found to employ uncertain parameters that significantly affect outputs. TopNet-GW benefitted from up-to-date calibrations and as such produced generally excellent simulations in comparison to observed streamflow. With the addition of the deep groundwater conceptual reservoir in the structure of the model, simulated flow baseflow index estimates and graphical assessment of flow recession curves indicate TopNet-GW reproduces groundwater processes well despite potential over-representation of baseflow at the expense of high flow periods during peak flows. These findings highlight the importance of combining subsurface and surface flow dynamics to resolve water management issues and improve model performance at the catchment scale.</p>

Exploring river–aquifer interactions and hydrological system response using baseflow separation, impulse response modelling and time series analysis in three temperate lowland catchments

Lowland rivers and shallow aquifers are closely coupled and their interactions are crucial for maintaining healthy stream ecological functions. In order to explore river–aquifer interactions and lowland hydrological system in three Belgian catchments, we apply a combined approach of baseflow separation, impulse response modelling and time series analysis over a 30–year study period at catchment scale. Baseflow from hydrograph separation shows that the three catchments are groundwater-dominated. The recursive digital filter methods generate a smoother baseflow time series than the graphical methods, and yield more reliable results than the graphical ones. Impulse response modelling is applied with a two–step procedure. The first step where groundwater level response is modelled shows that groundwater level in shallow aquifers reacts fast to the system input, with most of the wells reaching their peak response during the first day. There is an overall trend of faster response time and higher response magnitude in the wet (October–March) than the dry (April–September) periods. The second step of baseflow response modelling shows that the system response is also fast and that simulated baseflow can capture some variations but not the peaks of the separated baseflow time series. The time series analysis indicates that components such as interflow and overland flow, contribute significantly to stream flow. They are somehow included as part of the separated baseflow, which is likely to be overestimated from hydrograph separation. The impulse response modelling approach from the groundwater flow perspective can be an optional method to estimate the baseflow, since it considers some level of the physical connection between river and aquifer in the subsurface. Further research is however recommended to improve the simulation, such as giving more weight to wells close to the river and adding more drainage dynamics to the model input.

Can the two-parameter recursive digital filter baseflow separation method really be calibrated by the conductivity mass balance method?

The two-parameter recursive digital filter method (Eckhardt) and the conductivity mass balance (CMB) method are two widely used baseflow separation methods favored by hydrologists. Some divergences in the application of these two methods have emerged in recent years. Some scholars believe that deviation of baseflow separation results of the two methods is due to uncertainty of the parameters of the Eckhardt method and that the Eckhardt method should be corrected by reference to the CMB method. However, other scholars attribute the deviation to the fact that they contain different transient water components. This study aimed to resolve this disagreement by analyzing the effectiveness of the CMB method for correcting the Eckhardt method through application of the methods to 26 basins in the United States by comparison of the biases between the generated daily baseflow series. The results showed that the approach of calibrating the Eckhardt method against the CMB method provides a “false” calibration of total baseflow by offsetting the inherent biases in the baseflow sequences generated by the two methods. The baseflow sequence generated by the Eckhardt method usually includes slow interflow and bank storage return flow, whereas that of the CMB method usually includes high-conductivity water flushed from swamps and depressions by rainfall, but not low-conductivity interflow and bank storage return flow. This difference results in obvious peak misalignment and periodic deviation between the baseflow sequences obtained by the two methods, thereby preventing calibration. However, multi-component separation of streamflow can be achieved through comparison. Future research should recognize the deviations between the separation results obtained by the different methods, identify the reasons for these differences, and explore the hydrological information contained therein.

Comprehensive analysis and separation of river hydrograph using the GrWat R package

&lt;p&gt;River hydrograph analysis provides valuable information about temporal and spatial variability of the river discharge. One of the most imporant operations is separation of hydrograph, which aims at decomposing the total streamflow into components. Numerous approaches for hydrograph separation have been developed to date. Most of them traditionally separate the streamflow into general quickflow and baseflow components, but it is also possible to obtain more specific quickflow separation with subdivision into genetic components, such as seasonal snowmelt, rain, thaw etc. We present the general framework for river hydrograph analysis and separation provided by newly released GrWat package, which has been developed during several years. The framework includes a simple tabular data model for representation of hydrograph and climatic (temperature and precipitation) daily data needed for separation of the quickflow into genetic components; spatial analysis operations for automatic extraction of climatic data from reanalysis datasets covering the river basin; automated interpolation of missing data considering the autocorrelation; fast implementation of multiple algorithms for hydrograph separation; computation of more than 30 interannual and long-term characteristics of separated hydrograph components; scale-space transformation for hierarchical decomposition of the hydrograph; high-quality plotting and reporting of the results of analysis. One of the prominent features of the framework is a powerful algorithm for genetic hydrograph separation, which is capable of not only extracting the baseflow, seasonal, thaw and rain flood components, but also to cut the short-time rain floods which complicate the shape of the seasonal flood. The baseflow separation is performed on the first stage and can be initialized by any of the baseflow separation algorithms available in the package. On the second stage the quickflow is separated into genetic components. Such modular structure provides the flexible way to experiment with different combinations of algorithms and to select the approach wich serves best to the goal of the analysis and specific features of the hydrograph.&lt;/p&gt;&lt;p&gt;The study was supported by the Russian Science Foundation grant No. 19-77-10032&lt;/p&gt;

Evaluate recession characteristics from baseflow and separate baseflow via the knowledge guided machine learning

&lt;p&gt;Baseflow, referred to as the groundwater discharge, is essential to investigate the groundwater system. A common and classic approach to study baseflow is recession analysis method, but current methods confuse the concept of streamflow recessions and baseflow recessions. This confusion leads to a mixing effect of the fluxes from different storage components and theoretically inconsistent recession analysis results accordingly. Therefore, it motivates an improvement and enhanced scientific understanding of the empirically derived baseflow recession characteristics. &amp;#160;In addition, quantifying baseflow from streamflow is defined as the baseflow separation problem. The state-of-the-art baseflow separation tools are in lack of physical rules and have either structural limitations or are inapplicable in regions with insufficient data, which confines the generalization performance. To overcome these issues, we applied a knowledge guided machine learning (KGML) approach to separate baseflow, which embeds physically derived baseflow recession characteristics in the traditional machine learning framework.&lt;/p&gt;&lt;p&gt;Recession parameter, which is derived from empirical recession analysis, has been observed to exceed its theoretical range on a recession event scale. Besides many potential environmental factors, we hypothesize that this well recognized inconsistency is because the quick flow from surficial water bodies has not been successfully excluded based on the recession selection criterion. We conduct recession analysis using both streamflow and baseflow over 1,000 gages across the continental United States. The baseflow was estimated from Eckhardt two-parameter digital filter and was calibrated against the in-stream field data. It was found that for gages whose calibration performance is satisfactory, the baseflow derived recession parameter agrees more consistently with the recession characteristics, which are indicated by the Boussinesq solutions.&lt;/p&gt;&lt;p&gt;Traditional baseflow separation tools partition streamflow into quick flow and base flow. Those tools have data scarcity issues and structural limitations without involving physical perspectives. To introduce physical rules into baseflow separation and overcome data scarcity issues, we apply a recession-based loss function to train the machine learning model such that the recession characteristics of separated baseflow agree with their theoretical behaviors. Guided by the recession knowledge of baseflow on a catchment scale, progress is being made to finalize this KGML implementation and to improve the baseflow separation approach.&lt;/p&gt;

Modeling the Snowmelt Runoff Process of the Tizinafu River Basin, Northwest China, with GLDAS Data and Bayesian Uncertainty Analysis

The climate of the Tizinafu River basin is characterized by low temperature and sparse precipitation, and snow and glacier melt serve as the main water resource in this area. Modeling the snowmelt runoff process has great significance for local ecosystems and residents. The total streamflow of the Tizinafu River basin was divided into surface streamflow and baseflow. The surface streamflow was estimated using the routing model (RM) with Noah runoff data from Global Land Data Assimilation (GLDAS), and the parameter uncertainty of the RM was quantified through Markov chain Monte Carlo simulation. Additionally, the 10 commonly used baseflow separation methods of four categories [digital filter, hydrograph separation program (HYSEP), baseflow index, and Kalinlin methods] were used to generate the baseflow and were then evaluated by their performance in total streamflow simulation. The results demonstrated that the RM driven by GLDAS runoff data could reproduce the runoff process of the Tizinafu River basin. RM-Hl (local minimum HYSEP method) achieved the best performance in the total streamflow simulation, with Nash–Sutcliffe efficiency (NSE) coefficients of 0.82 and 0.93, relative errors of −0.40% and 10.50%, and observation inclusion ratios C of 62.07% and 68.52% for the calibration and verification periods, respectively. The local minimum HYSEP method was most suitable for describing the baseflow of the Tizinafu River basin among the 10 baseflow separation methods. However, digital filter methods exhibited weak performance in baseflow separation.