PosterPDF Available

Spatial allocation of low resolution runoff model outputs to high resolution stream network

  • April 2018
DOI:10.13140/RG.2.2.19122.09921
  • Conference: European Geosciences Union General Assembly 2018
Authors:
Marko Kallio at Aalto University
Marko Kallio
Joseph H. A. Guillaume at Australian National University
Joseph H. A. Guillaume
Matti Kummu at Aalto University
Matti Kummu
Fanuel Desalegn
Fanuel Desalegn
Fanuel Desalegn
  • This person is not on ResearchGate, or hasn't claimed this research yet.
Show all 5 authorsHide
Download file PDF
Read file
Download citation

Abstract

Global analyses of water scarcity and water resource management in developing countries would benefit from fine scale estimates of water availability at global extent. The standard procedure for hydrological predictions in ungauged basins includes 1) generation of a river network and watershed from a DEM, 2) derivation of meteorological forcing, either by interpolation of observations or downscaling from climate models, 3) estimation of parameters based on physical properties of the catchment. The procedure can produce detailed and relatively reliable information about water resources in catchments, but is non-trivial at global scale. On the other hand, predictions from global or continental scale models are widely available but have too low spatial resolution to describe precise conditions in the field. In this research, we test an approach to extract sub-grid resolution information by downscaling gridded, low resolution runoff products to an explicit high resolution river network.
Content uploaded by Marko Kallio
Author content
All content in this area was uploaded by Marko Kallio on May 10, 2018
Content may be subject to copyright.
Measurement station
HydroSheds 15 arc second river network
River segment Voronoi polygon
0.5 degree (30 min) grid
Country border
3S River Basin
107o
16o
106o
106.5o
107.5o
15o
15.5o
16o
15.5o
15o
14o
14.5o
13o
12.5o
12o
12o
14.5o
106o
106.5o
107o
107.5o
108o
108.5o
0
200
400 km
Mekong
3S
Laos
Vietnam
Cambodia
S
E
S
A
N
S
E
K
O
N
G
S
R
E
P
O
K
Station 440102
500
1000
1500
2000
Month
Avg Q [m3/s]
0
2 4 6 8 10 12
VISIT 30min
NRMSE: 14%
NSE: 0.98
R2: 0.98
Month
Avg Q [m3/s]
2 4 6 8 10 12
300
600
900
Station 430105
VISIT 30min
NRMSE: 21%
NSE: 0.95
R2: 0.97
0
1200
Station 430101
1000
2000
3000
4000
Month
Avg Q [m3/s]
0
2 4 6 8 10 12
VISIT 30min
NRMSE: 30%
NSE: 0.90
R2: 0.97
Month
Avg Q [m3/s]
2 4 6 8 10 12
500
1000
1500
LPJML 30min
NRMSE: 59%
NSE: 0.62
R2: 0.72
0
Station 450101
Equal weights
Strahler weights
Length weights
Voronoi basin
weights
DEM delineated
basin weights
Observed flow
Graph legend
Month
Avg Q [m3/s]
2 4 6 8 10 12
50
100
LPJmL 30min
NRMSE: 75%
NSE: 0.39
R2: 0.50
0
Station 440601
@ 30 minute
CARAIB, DBH, H08,
LPJ-GUESS, LPJML,
PCR-GLOBWB, MPIHM,
VIC, VISIT, WATERGAP,
WATERGAP2, WBM
List of models tested:
For full reference:
i.e. allocate weighted cell runoff to each river segment
compute_segment_runo( HSragrid or HSrgrid )
Runoff timeseries
( Raster or RasterBrick )
[ Area of interest ]
( polygon )
River network
( connected linestrings )
HSrgrid object
- weighted and routed river network
- HSgrid
HSragrid object
- routed river network
- weighted basin features
- HSgrid
compute_weights( river,HSgrid,
aoi],[basin],[drain.dir],[segment.weight] )
polygrid_timeseries( brick,[aoi] )
polygonized raster (HSgrid)
River network with flow timeseries ( HSflow )
accumulate_runo( HSrunoff )
using weights for
each river segment
by
using area of
catchment within a
grid cell to weight by
[ average_monthly_runoff( HSgrid ) ]
Analysis inputs
Function output
Legend
[ Optional ]
River network with segment specific runoff timeseries ( HSrunoff )
i.e. apply river routing
i.e. compute river segment weights by
a. river segment
Voronoi polygon
b. delineated from
drainage direction
c. user input
a. equal weights
b. segment length
c. Strahler stream
order
d. user input
Basin: River network:
Workflow in ‘hydrostreamer’ R package
1
2
3
4
Purpose
- To improve global water scarcity assessments.
- Runoff allocated to river segments within output grid cells (i.e.
downscaling runoff into explicit high-res river network).
- Done as simply and with the least input requirements as possible.
How? An Open Source R [1] package ‘hydrostreamer’
1. Create polygon grid from input raster
2. Weight river segments or basins within each grid cell
3. Assign grid cell value to river segments according to weights.
4. Apply river routing
- Minimum input data: runoff timeseries, river network
3S Basin Test Case
- 79 500 km2 tributaries of the Mekong - Sekong, Sesan and Srepok.
- Monsoon climate with distinct dry and wet season.
- Total runoff output from 12 models at 30 minute resolution obtained
from Inter-Sectoral Model Intercomparison Project (ISIMIP) [2]
- Tested also one model at 6 minute and another one with 3km
resolution.
- Simplest possible river routing: add everything downstream at each
timestep (month)
Results
- VISIT performs best at most stations
- Different weighting methods differ in results only at the smallest
streams. at higher stream orders the small differences upstream
are efficiently averaged out.
- When stream density-to-raster resolution gets too low, segment-
based weighting is not valid as not all cells contain river segments.
Conclusion and future ‘hydrostreamer’
- Results meaningful on monthly scale, but issues in the edges
of area of interest.
- Confirmed Karimipout et al [3] that Voronoi is viable alternative to
DEM delineated catchment areas.
- Recommended weighting by physical properties of segments:
either basin (Voronoi, or DEM delineated), or segment length.
- Investigate providing an interface in ‘hydrostreamer’ to existing river
routing applications (e.g. RAPID [4] or mizuRoute [5]).
- Add functions in ‘hydrostreamer’ to create optimal station-specific
model ensembles of several input models.
References
[1] R Core Team (2017). R: A language and environment for statistical computing. R Foundation for Statistical Computing, Vienna, Austria. URL https://www.R-project.org/.
[2] Potsdam Institute for Climate Impact Research. ISIMIP Data Archive. https://esg.pik-potsdam.de/projects/isimip/. Accessed: 28 March 2018.
[3] Karimipout, F. et al (2013). Watershed delineation from the medial axis of river networks, Computers & Geosciences, 59, 132-147. DOI: 10.1016/j.cageo.2013.06.004
[4] David, Cédric H. et al (2011). River network routing on the NHDPlus dataset, Journal of Hydrometeorology, 12(5), 913-934. DOI: 10.1175/2011JHM1345.1
[5] Mizukami, N. (2016). mizuRoute version 1: a river network routing tool for a continental domain water resources applications, Geoscientific Model Development, 9, 2223-2238. DOI: 10.5194/gmd-9-2223-2016
‘hydrostreamer’ in GitHub:
https://github.com/mkkallio/hydrostreamer
Spatial allocation of low resolution runo model outputs to
a high resolution stream network
Marko Kallio1*, Joseph H.A. Guillaume1, Fanuel Desalegn, Matti Kummu1, and Kirsi Virrantaus1
1 Department of Built Environment, Aalto University, Finland.
*Corresponding author: marko.k.kallio@aalto.
... Geheb et al. 2015). Kallio et al. (2018) have developed a tool in R programming language, called hydrostreamer, which downscales low-resolution runoff data to a high-resolution river network by assigning the runoff into river segments and accumulating flow from upstream to downstream, finally resulting in discharge estimates. The package can be used with openly available and global data, such as global runoff models, and the source code is freely available for download and modifications. ...
... One of the most notable sources for consistent and good quality near-global-scale river networks is HydroSHEDS (WWF 2018) that also provides drainage direction rasters 5 and pre-derived drainage basins that can serve as an area of interest. HydroSHEDS river network has been used by Kallio et al. (2018) during the development of hydrostreamer. However, since this study considers the spatial uncertainty related to the river network to be used, HydroSHEDS was applied only as benchmark data to compute discharge values to be compared with the simulation. ...
... Finally, two auxiliary data sets were employed: river lines describing the major rivers (Sekong, Sesan, Sre Pok) extracted from OpenStreetMap (OSM 2018) and the locations of HYMOS stations digitized by Kallio et al. (2018) as a part of hydrostreamer development work. The river network data was employed in the study because the uncertainty estimation required a reference that was based on an existing river network and to ensure that the river networks of each iteration will reach the HYMOS stations, in which it is absolutely known that there is a river since discharge measurements are available. ...
View
MizuRoute version 1: A river network routing tool for a continental domain water resources applications
Article
Full-text available
This paper describes the first version of a stand-alone runoff routing tool, mizuRoute. The mizuRoute tool post-processes runoff outputs from any distributed hydrologic model or land surface model to produce spatially distributed streamflow at various spatial scales from headwater basins to continental-wide river systems. The tool can utilize both traditional grid-based river network and vector-based river network data. Both types of river network include river segment lines and the associated drainage basin polygons, but the vector-based river network can represent finer-scale river lines than the grid-based network. Streamflow estimates at any desired location in the river network can be easily extracted from the output of mizuRoute. The routing process is simulated as two separate steps. First, hillslope routing is performed with a gamma-distribution-based unit-hydrograph to transport runoff from a hillslope to a catchment outlet. The second step is river channel routing, which is performed with one of two routing scheme options: (1) a kinematic wave tracking (KWT) routing procedure; and (2) an impulse response function – unit-hydrograph (IRF-UH) routing procedure. The mizuRoute tool also includes scripts (python, NetCDF operators) to pre-process spatial river network data. This paper demonstrates mizuRoute's capabilities to produce spatially distributed streamflow simulations based on river networks from the United States Geological Survey (USGS) Geospatial Fabric (GF) data set in which over 54 000 river segments and their contributing areas are mapped across the contiguous United States (CONUS). A brief analysis of model parameter sensitivity is also provided. The mizuRoute tool can assist model-based water resources assessments including studies of the impacts of climate change on streamflow.
View
Watershed Delineation from the Medial Axis of River Networks
Article
Full-text available
The watersheds are commonly delineated from digital elevation models (DEM). This approach is not efficient when an accurate DEM is not available. Furthermore, since raster-based algorithms are employed, the computations for large areas are very time consuming and even may be impractical. This article investigates delineation of the watersheds from the medial axis of river networks: If the river network is sampled by a set of points, the medial axis of the sample points provides an approximation of the catchments, whose aggregation results in the watersheds. Although this idea has been already proposed in the literature, the complexities of the medial axis extraction prevent it from being practically used. A major issue is appearing extraneous branches in the media axis due to perturbations of the sample points, which must be filtered out in a pre- or post-processing step. This article improves a Voronoi-based medial axis extraction algorithm by using labeled sample points to automatically avoid extraneous branches. The proposed method is used in four case studies to delineate the watersheds. The results illustrate that the proposed method is stable, easy to implement and robust, even in presence of significant noises and perturbations. The results also indicate that the watersheds delineated using the proposed and the DEM-based methods are reasonably comparable.
View
River network routing on the NHDPlus Dataset
Article
Full-text available
  • Oct 2011
The mapped rivers and streams of the contiguous United States are available in a geographic information system (GIS) dataset called National Hydrography Dataset Plus (NHDPlus). This hydrographic dataset has about 3 million river and water body reaches along with information on how they are connected into networks. The U.S. Geological Survey (USGS) National Water Information System (NWIS) provides streamflow observations at about 20 thousand gauges located on the NHDPlus river network.Ariver network model called Routing Application for Parallel Computation of Discharge (RAPID) is developed for the NHDPlus river network whose lateral inflow to the river network is calculated by a land surface model. A matrix-based version of the Muskingum method is developed herein, which RAPID uses to calculate flow and volume of water in all reaches of a river network with many thousands of reaches, including at ungauged locations. Gauges situated across river basins (not only at basin outlets) are used to automatically optimize the Muskingum parameters and to assess river flow computations, hence allowing the diagnosis of runoff computations provided by land surface models. RAPID is applied to the Guadalupe and San Antonio River basins in Texas, where flow wave celerities are estimated at multiple locations using 15-min data and can be reproduced reasonably with RAPID. This river model can be adapted for parallel computing and although the matrix method initially adds a large overhead, river flow results can be obtained faster than with the traditional Muskingum method when using a few processing cores, as demonstrated in a synthetic study using the upper Mississippi River basin.
View
View