Catchment Classification Research Articles

Identifying control factors of hydrological behavior through catchment classification in mainland China

Catchment classification based on hydrological similarity helps to understand the control factors of hydrological behavior. However, the relationship between hydrological behavior and its influencing factors has been unclear in mainland China because long-term and widely-distributed flow data is unavailable. Thus, this study intends to identify control factors of hydrological behavior in China’s basins by using classification. Gauged basins are clustered into several classes using the fuzzy c-means method based on flow signatures, which quantify catchment hydrological behavior. The classification and regression tree is employed to learn from cluster results and then obtain classes of basins without observed flow. Correlation methods are used to analyze the influence of basin signatures on flow signatures, while the difference significance test is applied to the hydrological behavior diversity between clusters from classification and regression tree. Results show that China’s basins are divided into five clusters, with low flow signatures more distinguishing classes than high flow signatures. It confirms that climate factors dominate hydrological behavior. However, soil is also an important control factor found in this study, which is rare in others. These findings help to understand hydrological behavior in China and reveal its control factors.

The Drivers of Hydrologic Behavior in Brazil: Insights From a Catchment Classification

AbstractDespite hosting ∼16% of the global freshwater and almost 50% of water resources in South America, Brazilian catchment‐scale relationships between drivers and streamflow are still poorly understood. Here, we used streamflow signatures and attributes of 735 catchments from the Catchment Attributes for Brazil data set to investigate the dominant hydrological processes for the catchments. We also assess how catchments group based on hydrologic behavior similarities and analyze which climatic/landscape attributes control the streamflow variability. To classify and group the catchments, we used the k‐means method optimized by the Elbow approach, along with a Principal Component Analysis. Uncertainty on catchment grouping was checked by k‐fold cross‐validation. Then, we used a recursive feature elimination using the random forest technique to assess the most influential catchment attributes to the hydrological signatures. Our results revealed six similarity groups, which followed mainly an aridity gradient ranging from the wettest to the driest, but also seasonality. The climate is the primary driver of hydrological behavior for the water‐limited groups, highlighting the influence and importance of the atmospheric demand in several Brazilian catchments. High soil storage capacity in energy‐limited catchments associated with high precipitation led to high discharge all year due to the subsurface fluxes' contribution. Our findings may be useful to improve streamflow predictability and hydrological behavior identification by further understanding hydrological similarities and their signatures due to catchment landscape characteristics. Further, by employing an easily reproducible methodology and clear metrics to weigh uncertainty, our study provides a significant step toward establishing a catchment‐scale common classification system.

Streamflow prediction in ungauged catchments through use of catchment classification and deep learning

Streamflow prediction in ungauged catchments is a challenging task in hydrological studies. Recently, data-driven models have demonstrated their superiority over traditional hydrological models in predicting streamflow in ungauged catchments. However, previous studies have overlooked the similarities between the training and the target catchments. Therefore, this study explores the role of catchment similarity in regionalization modeling using the publicly available CAMELS dataset. We employed the dynamic time warping-based KMeans (DTW-KMeans) time-series clustering technique to cluster the streamflow data from gauged catchments. We utilized the long short-term memory (LSTM) neural network to construct regional models for different classes of gauged catchment. Additionally, the mapping relationship between gauged catchment classes and static attributes was established using the random forest (RF). By combining the trained RF model with the static attributes of an ungauged catchment, we determined its class and used the corresponding regional LSTM to predict streamflow. To evaluate the effectiveness of the framework, we applied the classification-based regionalization modeling (CRM) and non-classification-based regionalization modeling (NRM) approach for comparison. The results indicate that: (1) The DTW-KMeans-based catchment classification method is generally accurate and reasonable; (2) the complexity of the LSTM model and the number of training catchments should be appropriately matched to improve streamflow prediction; and (3) catchment similarity plays a crucial role in regionalization modeling, the proportion of training catchments with high similarity to ungauged catchments significantly affects prediction results.

Modularity Density Based-Edge Betweenness Method for Catchment Classification in Sabah

One of the most significant studies in hydrology is the classification of catchments. There are various reasons why catchment classification studies have been conducted, but most importantly is to anticipate the prediction of ungauged basin (PUB), among others. There are several ways of classification, each with its own set of assumptions and basis, that have been used for catchment classification. Recently, complex network concepts, specifically community structure, have emerged as key classification methods, and are now attaining traction in catchment classification. As a result, in this study, community structure approach, specifically the Modularity Density based- Edge Betweenness (MDEB) method is implemented to split the network into communities. By using the proposed method, a network of 30 monthly streamflow stations across Sabah is considered for catchment classification. The impact of correlation threshold, which vary from 0 to 1, denoting the strength between a pair of catchments is also investigated. As a result, four threshold values are chosen (T = 0.5, 0.55, 0.6, and 0.65), and the communities established with threshold value, T=0.6 are interpreted. The relationship between catchment characteristics and flow characteristics are investigated as well as distance-correlation relationships for communities identified to understand the Sabah catchment’s behaviors.

Community Detection Using Modularity Optimization Method For Catchment Classification

A general framework for catchment classification may be helpful for more accurate and efficient modeling of hydrologic systems, as well as to improve communication between hydrology researchers and those in other disciplines. There are plethora numbers of methods applied for catchment classification, but in these years, recent studies are implementing the complex networks concept for classification purposes. The community structure methods which are complex networks-based methods are focus mainly to classify catchments. Hence, the efficiency of complex network ideas, especially using the methods of community structure is examined in this study. Specifically, the modularity optimization method that is one of the community structure methods is applied to classify 218 stream-gauges stations in entire Australia that covers a large variety of hydroclimatic, topographic, geomorphic, soil usage, and climatic parameters. In the present study, the applicability and the efficiency of the community structure concept is validated by the proposed method. The classification of Australian catchments was further assessed with threshold value of 0.8, which resulted formation of nine communities with at least 9 stations in a community which combine to have almost 77% of the total number of stations (165 out of 218). All nine selected communities were also examined in terms of the flow characteristics (i.e. flow mean and flow covariance) and the catchment characteristics (i.e. drainage area, elevation and stream length). The catchment behaviors for each selected communities were also interpreted in terms of distance and correlation relationship, which give some useful insights towards generalization of hydrologic framework.

A network approach for multiscale catchment classification using traits

Abstract. The classification of river catchments into groups with similar biophysical characteristics is useful to understand and predict their hydrological behavior. The increasing availability of remote sensing and other large-scale geospatial datasets has enabled the use of advanced data-driven approaches to classify catchments using traits such as topography, geology, climate, land cover, land use, and human influence. Unsupervised clustering algorithms based on the Euclidean distance are commonly used for trait-based classification but are not suitable for highly dimensional data. In this study we present a new network-based method for multi-scale catchment classification, which can be applied to large datasets and used to determine the traits associated with different catchment groups. In this framework, two networks are analyzed in parallel: the first being where the nodes are traits and the second being where the nodes are catchments. In both cases, edges represent pairwise similarity, and a network cluster detection algorithm is used for the classification. The trait network is used to investigate redundancy in the trait data and to condense this information into a small number of interpretable categories. The catchments network is used to classify the catchments into clusters and to identify representative catchments for the different groups using the degree centrality metric. We apply this method to classify 9067 river catchments across the contiguous United States at both regional and continental scales using 274 non-categorical traits. At the continental scale, we identify 25 interpretable trait categories and 34 catchment clusters of sizes greater than 50. We find that catchments with similar trait categories are typically located in the same region, with different spatial patterns emerging among clusters dominated by natural and anthropogenic traits. We also find that the catchment clusters exhibit distinct hydrological behavior based on an analysis of streamflow indices. This network approach provides several advantages over traditional means of classification, including better separation of clusters, the use of alternate similarity metrics that are more suitable for highly dimensional data, and reducing redundancy in the trait information. The paired catchment–trait networks enable analysis of hydrological behavior using the dominant trait categories for each catchment cluster. The approach can be used at multiple spatial scales since the network topologies adjust automatically to reflect the trait patterns at the scale of investigation. Finally, the representative catchments identified as hub nodes in the network can be used to guide transferable observational and modeling strategies. The method is broadly applicable beyond hydrology for classification of other complex systems that utilize different types of trait datasets.

A Nonlinear Local Approximation Approach for Catchment Classification.

Catchment classification plays an important role in many applications associated with water resources and environment. In recent years, several studies have applied the concepts of nonlinear dynamics and chaos for catchment classification, mainly using dimensionality measures. The present study explores prediction as a measure for catchment classification, through application of a nonlinear local approximation prediction method. The method uses the concept of phase-space reconstruction of a time series to represent the underlying system dynamics and identifies nearest neighbors in the phase space for system evolution and prediction. The prediction accuracy measures, as well as the optimum values of the parameters involved in the method (e.g., phase space or embedding dimension, number of neighbors), are used for classification. For implementation, the method is applied to daily streamflow data from 218 catchments in Australia, and predictions are made for different embedding dimensions and number of neighbors. The prediction results suggest that phase-space reconstruction using streamflow alone can provide good predictions. The results also indicate that better predictions are achieved for lower embedding dimensions and smaller numbers of neighbors, suggesting possible low dimensionality of the streamflow dynamics. The classification results based on prediction accuracy are found to be useful for identification of regions/stations with higher predictability, which has important implications for interpolation or extrapolation of streamflow data.

Spatial Variability of Rainfall and Classification of Peninsular Indian Catchments

The strength and success of hydrological analysis depend upon the quantity and quality of observed data. In the recent past, the availability of advanced computing facilities and measurement techniques had a great impact on the field of hydrology, especially in hydrologic analysis and hydrologic modeling. In spite of such growth, the present hydrologic modeling has certain challenges: complexity (involving a large number of parameters), applicability to a specific region (difficult to generalize for other regions), and lack of understanding of the connection between model theories and the actual system. The general solution of simplifying the models in terms of developing a classification framework has been discussed and focused on in the present study. It will greatly help to overcome the hydrologic modeling challenges and provides a better understanding of the hydrologic process. In general, classification is a way of grouping entities which has similar characteristics. The importance of applying nonlinear dynamics and chaos methods for classification has been realized in the recent past; since such studies provide exclusive information on hidden characteristics such as complexity, nonlinearity, dimensionality, etc. Of hydrological processes. The hydrologic processes are complex. In this study, information regarding the complexity is extracted by statistical analysis and linear methods such as Autocorrelation Function, and Average Mutual Information. 367 gridded rainfall stations over Peninsular Indian basins are used to investigate the applicability of different methods used in the study.

Control of climate and physiography on runoff response behavior through use of catchment classification and machine learning

Understanding of runoff response changes (RRC) is essential for water resource management decisions. However, there is a limited understanding of the effects of climate and landscape properties on RRC behavior. This study explored RRC behavior across controls and predictability in 1003 catchments in the contiguous United States (CONUS) using catchment classification and machine learning. Over 1000+ catchments are grouped into ten classes with similar hydrological behavior across CONUS. Indices quantifying RRC were constructed and then predicted within each class of the 10 classes and over the entire1000+ catchments using two machine learning models (random forest and CUBIST) based on 56 indicators of catchment attributes (CA) and 16 flow signatures (FS). This enabled the ranking of the important influential factors on RRC. We found that (i) CA/FS-based clusters followed the ecoregions over CONUS, and the impact of climate on RRC seemed to overlap with physiographic attributes; (ii) CUBIST outperforms the random forest model both within the cluster and over the whole domain, with a mean improvement of 39 % (depending on clusters) within clusters. Runoff sensitivity was better predicted than runoff changes; (iii) FS related to runoff ratio, average, and high flow are the most important for RRC, whereas climate (evaporation and aridity) is a secondary factor; and (iv) RRC patterns are substantial in the dominant factor space. High total changes and catchment characteristic-induced changes occurred mainly at 100°west longitude. The elasticity of climate and catchment characteristics was found to be high in spaces with high evaporation and low runoff ratios and low in spaces with low evaporation and high runoff ratios. Uncertainties existed in the number of catchments between clusters which was verified using a fuzzy clustering algorithm. We recommend that future research that clarifies the impact of uncertainty on hydrological or catchment behavior should be conducted.

Changes in the hydrological and sediment regimes of two neighboring catchments in the past sixty years

Despite advances in watershed hydrology research, little is known about the hydrologic differences and similarities between neighboring catchments. Two neighboring catchments, the Nanliu River catchment (NLRC) and the Qinjiang River catchment (QJRC), were studied to see how they differed in annual precipitation, runoff, and sediment load in the last 60 years. Four hydrological signatures (river regime, periodicity, change trends, and precipitation-runoff-sediment correlations) estimated from a wide variety of catchment classification and corresponding indices, such as the runoff ratio (RQP) and baseflow index (BFI). The findings revealed that the two catchments shared similar general features, including similar runoff patterns (RQP= 0.45 and 0.43, and BFI = 2.83% and 2.12% in NLRC and QJRC respectively), minor differences in periodicities, and similar annual precipitation, runoff, and sediment load changes. The annual sediment load has been steadily decreasing due to human activity. The simple precipitation-runoff regression model developed in one catchment can be transferred to another. Nash–Sutcliffe efficiency (NSE) coefficients were 0.79 and 0.62, respectively. The runoff-sediment load regression model, on the other hand, showed poor performance (NSE of −0.34 and −0.99, respectively), despite capturing the main patterns with correlation coefficients of 0.82 (p < 0.001) and 0.74 (p < 0.001), respectively. These findings suggested that (1) annual runoff ratios are influenced more by climate (particularly precipitation) than by surface and subsurface characteristics of the earth. The rainfall-runoff model for the neighboring catchment can recreate runoff dynamics at yearly scales. Human activities, such as dam construction and sand mining, significantly impacted the annual sediment yield. (2) The two neighboring catchments had similar hydrologic ages because of similar climate, geology, tectonics, and human disturbance. These findings provided a solid evidence for predicting runoff and sediment in ungauged basins.

Pattern recognition describing spatio-temporal drivers of catchment classification for water quality

Classification using spatial data is foundational for hydrological modelling, particularly for ungauged areas. However, models developed from classified land use drivers deliver inconsistent water quality results for the same land uses and hinder decision-making guided by those models. This paper explores whether the temporal variation of water quality drivers, such as season and flow, influence inconsistency in the classification, and whether variability is captured in spatial datasets that include original vegetation to represent the variability of biotic responses in areas mapped with the same land use. An Artificial Neural Network Pattern Recognition (ANN-PR) method is used to match catchments by Dissolved Inorganic Nitrogen (DIN) patterns in water quality datasets partitioned into Wet vs Dry Seasons and Increasing vs Retreating flows. Explainable artificial intelligence approaches are then used to classify catchments via spatial feature datasets for each catchment. Catchments matched for sharing patterns in both spatial data and DIN datasets were corroborated and the benefit of partitioning the observed DIN dataset evaluated using Kruskal Wallis method. The highest corroboration rates for spatial data classification with DIN classification were achieved with seasonal partitioning of water quality datasets and significant independence (p < 0.001 to 0.026) from non-partitioned datasets was achieved. This study demonstrated that DIN patterns fall into three categories suited to classification under differing temporal scales with corresponding vegetation types as the indicators. Categories 1 and 3 included dominance of woodlands in their datasets and catchments suited to classify together change depending on temporal scale of the data. Category 2 catchments were dominated by vineforest and classified catchments did not change under different temporal scales. This demonstrates that including original vegetation as a proxy for differences in DIN patterns will help guide future classification where only spatially mapped data is available for ungauged catchments and will better inform data needs for water modelling.

Identifying Hydrologic Regimes and Drivers in Nova Scotia, Canada: Catchment Classification Efforts for a Data-Limited Region

Identifying Hydrologic Regimes and Drivers in Nova Scotia, Canada: Catchment Classification Efforts for a Data-Limited Region

Hydrological characteristics of Australia: national catchment classification and regional relationships

• Australian subcatchments are described by forty climate and intrinsic attributes. • Fuzzy c-means classified the national subcatchment dataset into eight clusters. • Most subcatchments have a very high membership coefficient to one cluster. • Each cluster is typified by specific streamflow behaviour and flood risk. • Subcatchments with poor cluster membership levels highlight interactive complexity. A continent-wide classification study of Australian catchments was undertaken to group drainage basins with comparable flow characteristics based on forty metrics from the themes of climate, topography, surface condition and hydrogeology. A principal component analysis reduced the multi-collinear indices to nine principal components (PCs). A fuzzy c-means cluster analysis on the PCs resulted in the delineation of eight subcatchment clusters with similar physiography, and 82.5% of subcatchments displaying a membership coefficient greater than 0.7. Subcatchments with poor cluster membership levels may be significant in highlighting complexity in the interaction between climatic controls and local geomorphological and land use variability. The clusters are distinguished by different attributes in several key subcatchment variables and each group is described in terms of its flow and flood profile. These interpretations are supported by separately derived streamflow signatures from gauged catchments. Cluster and PC distribution mapping revealed distinct spatial and drainage relationships between the subcatchments relevant to the assessment of flood behaviour. The spatial pattern of the river basin regionalisation via physical catchment similarity closely resembles the modified Köppen climate classification, with differences highlighting the importance of non-climate characteristics on streamflow behaviour in Australia. The catchment classification is an Australia-wide regionalisation scheme, incorporating datasets with disparate spatial distribution and ungauged subcatchments. This study therefore provides the framework for more detailed streamflow analysis by applying consistent methods across all catchments in Australia, and has implications for flood risk prediction, mitigation and planning.

Exploring Homogeneity among Catchments forEfficient Province-wide Watershed Managementin Negros Occidental, Philippines

Catchment classification is one approach in natural resource management that is widely adopted in taking efficient steps towards implementing suitable soil and water conservation measures across a basin or region. Catchments have unique characteristics emerging from the heterogeneity and complexity of the systems and classifying them paves way to achieve order and simplicity. However, some constraints related to data availability could be a problem in a region where only few rivers are gauged and with only one type of climate data available. This study presents a way to decrease complexity by grouping these catchments based on their biophysical characteristics extracted from readily available datasets and using simple statistical approaches. Principal component analysis was first conducted to twenty-four biophysical variables which were reduced to eight factor components. A hierarchical clustering method was then performed to define the number of clusters and K-means clustering procedure was followed for the final grouping. Nine watershed clusters were formed with watershed size having the greatest contribution. Grouping catchments into clusters with similar biophysical characteristics does not only promote simplicity but also facilitates understanding of the nature of not only one watershed but also its relationship with other watersheds in a bigger landscape. The study also confirmed that spatially close watersheds exhibit similar characteristics.

Incorporating Landscape Scaling Relations into Catchment Classification for Optimizing Ecological Management

The landscape scaling relation challenges catchment ecological management; however, how the scaling relations change among naturally and anthropogenically differentiated catchments is still unknown. In this study, approximately 1500 soil samples were determined; more than 800 households were surveyed; and the landscape pattern was investigated in 120 sub-catchments of a subtropical Chinese urbanizing agricultural catchment. A scalogram and a coefficient of variation of the commonly used landscape metrics were estimated among various grain sizes, to quantify the Strength of Landscape Scale Effects (SSE) among sub-catchments. Natural and anthropogenic determinants for the SSE were determined. Then, the determinants incorporating landscape scaling relation were applied to classify the sub-catchments through the k-means clustering analysis. The SSE presented different spatial heterogeneity across the 120 sub-catchments and was not expectedly related to the scaling relation over the entire catchment, especially for the Contagion index and Shannon’s Evenness Index. The SSE were significantly related to natural and anthropogenic factors including the soil sand content, the population density, the relief ratio, and the ratio of arable land to woodland. The four factors combing with landscape scaling relations contributed to the four gratifying convergent categories for the 120 sub-catchments. Category I with a large relief and less anthropogenic disturbance had higher spatially non-stationary relationship, while categories II, III, and IV, with varying degrees of relatively small relief and strong intensities of anthropogenic disturbance, had a lower spatial heterogeneity of the landscape scaling relation. The results implied that category I was required to strengthen environmental protection of spatial differences, and categories II, III, and IV could ignore the landscape scale effects and even upscaling management to save management resources when carrying out ecological management within. Our findings could minimize uncertainty in ecological planning and provide opportunities for the application of multiple-scale management.

골목상권의 영역 설정과 유형 분류에 관한 연구: 청주시를 사례로

Alley business districts are important market areas to secure retail and commercial marketing competitiveness and stability, most of which has been vulnerable to external shocks, including the COVID-19 pandemic. While, most research studies in South Korea have focused on cases within large cities and high-street business districts such as Seoul and several large-sized cities in Kyuunggi province, few studies have attempted to delineate alley business district on a set of street boundaries in small and medium-sized cities where one-person business owners and self-employed subcontractors, in particular, have been suffered. This study has two main objectives: delineating spatial extents and classifying district types of the alley business districts suitable for small and medium-sized cities in Korea. This study delineates spatial extents using GIS-based data handling and processing techniques with the data sets of individual store coordinates and ‘Gil’ type of the road names available through small enterprise, commercial, and road name address databases. Using the delineating spatial extent and classification results, this study produced alley business districts map of Cheongju city and examined the spatial characteristics of the four types of districts in the city. Finally, we expect that the findings of this study — alley(‘Gil’)-based catchment areas and classification of the alley businesses — may be used as core data and contribute adequately to predicting street-scaled markets and businesses. Besides, the results of this study can serve as a realistic starting point for policymaking to revitalize alley business districts with the understanding and prediction of its district changes.

Investigating hydrologic controls on 26 Precambrian shield catchments using landscape, isotope tracer and flow metrics

AbstractAlthough many studies focus on catchment classification, hydrologists are challenged to find the most important variables for a meaningful catchment classification, especially in the heterogeneous environment of the Precambrian Shield. This study investigates landscape controls on hydrologic response by conducting a catchment classification of 26 catchments located within two Precambrian Shield watersheds (Sturgeon–French–Nipissing [SNF] and Muskoka) in central and Northeastern Ontario, Canada, using combinations of landscape characteristics (e.g. topography, geology, landcover), hydrometric variables, and stable isotopes of δ18O and δ2H in river flow. Weekly to monthly surveys of δ18O and δ2H in river flow were collected between 2013 and 2019. Flow metrics (e.g. Pardé coefficient and coefficient of variation of streamflow) were generated for 14 of the 26 rivers between 2008 and 2018. Principal component analyses (PCA) and Hierarchical Clustering on Principal Components (HCPC) analyses were used to identify variables controlling catchment clustering according to their similarities for four different scenarios. Despite their similar location along the southern edge of the Precambrian Shield, the 26 catchments generally clustered by watershed (SNF and Muskoka) with some exceptions. Differences in wetland and lake area (%), mean slope, and % area covered by glacialacustrine and glaciofluvial outwash deposits were the most influential variables in catchment classification. A positive correlation between % wetland area and streamflow stable isotope damping ratios suggests greater % wetland area (with shallow and potentially seasonally variable surface areas and/or hydrologic connection) observed in the SNF catchments increases variability in the influence of evaporative enrichment in SNF catchments. The catchment classification analyses in combination with stable isotopes of δ18O and δ2H were functional tools to investigate the combined influences of diverse types of catchment characteristics that lead to differences in hydrometric response. These results could support future studies focusing on generating hydrologic models and representing wetlands connectivity in the region.

Classification of Trichonis Lake graben (Western Greece) alluvial fans and catchments using geomorphometry and artificial intelligence

Classification of Trichonis Lake graben (Western Greece) alluvial fans and catchments using geomorphometry and artificial intelligence

Classifying annual daily hydrographs in Western North America using t‐distributed stochastic neighbour embedding

AbstractFlow regimes are critical for determining physical and biological processes in rivers and their classification and regionalization traditionally seeks to link patterns of flow to physiographic, climate and other information. There are many approaches to, and rationales for, catchment classification, with those focused on streamflow often seeking to relate a particular response characteristic to a physical property or climatic driver. Rationales include such topics as prediction in ungauged basins (PUB), and providing guidance for model selection in poorly understood hydrological systems. The annual daily hydrograph (ADH) is a first‐order easily visualized integrated expression of catchment function, and over many years the average ADH is a distinct hydrological signature that differentiate catchments from each other. In this study, we use t‐SNE, a state‐of‐the‐art technique of dimensionality reduction, to classify 17 110 ADHs for 304 reference catchments in mountainous Western North America. t‐SNE is chosen over other conventional methods of dimensionality reduction (e.g., PCA) as it presents greater separability of ADHs, which are projected on a 2D map where the similarities are evaluated according to their map distance. We then utilize a Deep Learning encoder to upgrade the non‐parametric t‐SNE to a parametric approach, enhancing its capability to address ‘unseen’ samples. Results showed that t‐SNE successfully clustered ADHs of similar flow regimes on the 2D map and allowed more accurate classification with KNN. In addition, many compact clusters on the 2D map in the coastal Pacific Northwest suggest information redundancy in the local streamflow network. The t‐SNE map provides an intuitive way to visualize the similarity of high‐dimensional data of ADHs, groups catchments with like characteristics, and avoids the reliance on subjective hydrometric indicators.

Classification of catchments for nitrogen using Artificial Neural Network Pattern Recognition and spatial data

In hydrological modelling, classification of catchments is a fundamental task for overcoming deficits in observational datasets. Most attention on this issue has focussed on identifying the catchments with similar hydrological responses for streamflow. Yet, effective methods for catchment classification are currently lacking in respect to Dissolved Inorganic Nitrogen (DIN), a water quality constituent that, at increasing concentrations, is threatening nutrient sensitive environments. Pattern recognition, using standard Artificial Neural Network algorithm is applied, as a novel approach to classify datasets that are considered to be suitable proxies for biological and anthropogenic drivers of observed DIN releases. Eleven gauged Great Barrier Reef (GBR) catchments within Queensland Australia are classified using spatial datasets extracted from ecosystem (e.g. original ecosystem responses to biogeographic, land zone, land form, and soil type attributes) and land use maps. To evaluate the performance of the examined spatial datasets as a proxy for deductive classification, the classification process is repeated inductively, using observed DIN and streamflow data from gauging stations. The ANN-PR method is seen to generate the same classification score format for the differing dataset types, and this facilitates a direct comparison for model output for observed data corroborations. The Kruskal-Wallis test for independence, at p > 0.05, identifies the deductive classification approach as a predictor for classification using DIN observations, which lacks an independence from each other at a p value of 0.01 and 0.02. This study concludes that an ANN-PR method can integrate the ecosystem and land use mapping data to deductively classify the GBR catchments into four regions that also have similar patterns of DIN concentrations. Due to the uniform availability of the mapping data, the findings provide a sound basis for further investigations into the transposing of knowledge from gauged catchments to ungauged areas.