Hydroinformatics advances for operational river forecasting: using graphs for drainage network descriptions

Distributed hydrologic models provide accurate river streamflow forecasts and a multitude of spatially varied products on basin scales. The distributed elements of the basins are pieced together using drainage networks. An efficient representation of drainage networks in computer code is necessary. Graph theory has long been applied in many engineering areas to solve network problems. In this paper we demonstrate that adjacent list graph is the most efficient way of presenting the drainage network in terms of development and execution. The authors have implemented drainage networks using the adjacency-list structure in both the research and operational versions of the US National Weather Service (NWS) distributed model. A parallel routing algorithm based on Dijsktra's shortest path algorithm was also developed using the MPI library, which was tested on a cluster using the Oklahoma Illinois River basin dataset. Theoretical analysis and test results show that inter-processor communication and unbalanced workload among the processors limit the scalability of the parallel algorithm. The parallel algorithm is more applicable to computers with high inter-processor bandwidth, and to basins where the number of grid cells is large and the maximum distance of the grid cells to the outlet is short.

[1]  S. K. Jenson,et al.  Extracting topographic structure from digital elevation data for geographic information-system analysis , 1988 .

[2]  M. Brilly,et al.  Automated grid element ordering for GIS-based overland flow modeling , 1992 .

[3]  Timothy H. Keitt,et al.  LANDSCAPE CONNECTIVITY: A GRAPH‐THEORETIC PERSPECTIVE , 2001 .

[4]  Henry Neeman,et al.  Parallelisation of a distributed hydrologic model , 2005, Int. J. Comput. Appl. Technol..

[5]  R. L. Shreve.  Infinite Topologically Random Channel Networks , 1967, The Journal of Geology.

[6]  Francesco Maffioli,et al.  Graphs and combinatorial optimization , 2006, Discret. Optim..

[7]  T. Hamill.  The National Weather Service River Forecast System , 1999 .

[8]  Richard T. T. Forman,et al.  Landscape graphs: Ecological modeling with graph theory to detect configurations common to diverse landscapes , 1993, Landscape Ecology.

[9]  Victor Koren,et al.  A Satellite Based River Forecast System for the Nile River , 1994 .

[10]  Use of Topologic Information in Processing Data for Channel Networks , 1970 .

[11]  John F. O'Callaghan,et al.  The extraction of drainage networks from digital elevation data , 1984, Comput. Vis. Graph. Image Process..

[12]  Boris Schröder,et al.  Pattern, process, and function in landscape ecology and catchment hydrology – how can quantitative landscape ecology support predictions in ungauged basins? , 2006 .

[13]  Michael Smith,et al.  Hydrology laboratory research modeling system (HL-RMS) of the US national weather service , 2004 .

[14]  R. K. Shyamasundar,et al.  Introduction to algorithms , 1996 .

[15]  Rajiv Gupta,et al.  Extended Use of Linear Graph Theory for Analysis of Pipe Networks , 2000 .

[16]  Seann Reed,et al.  Deriving flow directions for coarse‐resolution (1–4 km) gridded hydrologic modeling , 2003 .

[17]  J. Obeysekera,et al.  Graph Theory Data Objects Applied to Stream Flow Network Representation in an Integrated Hydrological Model , 2005 .

[18]  Dong-Jun Seo,et al.  The distributed model intercomparison project (DMIP): Motivation and experiment design , 2004 .

[19]  Philippe Lagacherie,et al.  Modelling Spatial Variability Along Drainage Networks with Geostatistics , 2006 .

[20]  Bin Jiang,et al.  A Structural Approach to the Model Generalization of an Urban Street Network* , 2004, GeoInformatica.

[21]  J. Fairfield,et al.  Drainage networks from grid digital elevation models , 1991 .

[22]  Jean-Yves Berthou,et al.  Comparing OpenMP, HPF, AND MPI Programming: A Study Case , 2001, Int. J. High Perform. Comput. Appl..

[23]  Theodore K. Apostolopoulos,et al.  Parallel computation for streamflow prediction with distributed hydrologic models , 1997 .

[24]  Herb Sutter,et al.  A Fundamental Turn Toward Concurrency in Software , 2008 .

[25]  Alon Rimmer,et al.  A FAST RECURSIVE GIS ALGORITHM FOR COMPUTING STRAHLER STREAM ORDER IN BRAIDED AND NONBRAIDED NETWORKS 1 , 2004 .

[26]  Pavel Kabat,et al.  Integrating hydrology, ecosystem dynamics, and biogeochemistry in complex landscapes , 1999 .

[27]  D. M. Coffman,et al.  Computer Determination of the Geometry and Topology of Stream Networks , 1971 .

[28]  A. E. Scheidegger,et al.  On the topology of river nets , 1967 .

[29]  Jeremy G. Siek,et al.  The generic graph component library , 1999, OOPSLA '99.