GIS solution in defining a basin for a choke point along a drainage line

Presenter : Praveen Kumar Sinha

Type: Oral

Devastation by excessive precipitation in metropolises like Mumbai, Chennai and Bangalore during the Rainy Season 2005 is a cause of concern to society. Flash floods followed by heavy rains generated panicky among the dwellers. Why has this occurred in a particular street or a specific patch of a city during the Rainy Season-2005 or 2016? Recursive floods in other parts of India are also a point of grave concern. Over one hundred people traveling in a train across Nalagonda district, in close proximity of Hyderabad, were washed away in flash floods of 2015. Causes to these events were of course heavy rains in the basin. But human interference in nature is also equally responsible. This warrants a study of basin, in the concerned region.

Dimensional extent of a basin, elevation distribution all-through basin, detailed drainage line distribution, hydrological behavior of the group of rocks occurring in a basin, rain gauge – reports from the basin, siltation along the drainage line, human interference by creating impervious concrete structures, deforestation and pouring the flow choking wastes in the drainages of a basin etc. are the thrust areas to come to any study based conclusion or speak about any safety measures. Study based report or map of hazard / risk-zonation within a basin is the specific requirement for any basin management team. Studies focused within a basin on different parameters as causative to flood, is obvious, to come to any scientific conclusion. At the time of urgencies, a basin for a point along a drainage line is of instantaneous importance – to carry out an analysis of instantaneous nature.

Information Technology (IT) is a very useful tool for speedy retrieval and processing of voluminous data to arrive at new meaningful interpretation, request and requirement based processed spatial and application based information dissemination to cater the needs of planning, forcasting, management and discision support activities for the general benefits of public. Geographic Information Systems (GIS) are the brain child of the fast growing technology to carry out any analysis based on spatial database. Costly and high-performance softwares implementing object-based or topology implemented connectivity-defined detailed-drainage-network are available in the market which could select the connected chain of lines above(MSL) a point – may be a point of causing congestion, flume station, dam site along the flow regime. And so is to define the basin zone around the selected and sub-setted drainage line for all the water drops falling in the area of influence tend to rush down slope to the point of concern (choking point – causing flood to upstream side).

This paper is on the combined approach with vector and raster GIS implementations in defining a basin above such a point of concern. Implementation of defining a basin in a Free and Open Source GIS – GRASS or ILWIS. Conceptually, and fundamentally this software is not different from any costly GIS. Basin implementation at a point on a drainage line may requre three steps:

  • Developing a node at the choke point, selecting and subsetting the connected network of streams above the node.
  • Development of two separate distance rasters:
    1. around the drainage line database : with resulting output raster, to say named – DISTarndDR, a continuous raster for distance for the extent of whole study area  and
    2. around the subsetted drainage lines: with resulting output raster, to say named –    DISTarndSEL_DR, a continuous raster for distance for the extent of extent of study area.
  • Making a conditional (if, then and else statement) statement (ILWIS) of a nature as below:

BASIN=IFF(DISTarndSEL_DR = DISTarndDR, 1, null).

The fundamentals lies with computation of distane in raster GIS.

In raster GIS, nearest neighbourhood as a function of distance is calculated by a combination of 4-cells and 8-cells-connected dilation or spread around the predictor cells. Spread computation start taking place around one or more predictor locations, or the source locations as they are the locations of the source of the spread of distances. From the source locations the GIS will be able to compute a new raster that will indicate how much minimal total distance the spread has witnessed for reaching a raster cell. The spread from a source cell (Csrc ) to the nearest cell, i.e., the cell may be to the E, W, N or S are closest neighbour than to the cell that is to its NW, NE, SE or SW. The distance ratio between these two cases is 1:Ö2.   The distance calculation is a process in which for each pixel the distance to its neighboring pixels is calculated using a 3´3 matrix with the type shown here:

7 5 7
5 0 5
7 5 7

7/5 is a good approximation of Ö2, and Ö2 is the distance between two diagonally connected pixels when the raster cell size (or resolution) is 1.

Spread of distances from the predictor cells is a recursive (reiterative) process in which the output raster map has to be scanned forward and backward until no more changes occur. The final result of the operation is a radial distance raster grown from the predictor cells. For each pixel (column by column, line by line) location which is located as a diagonal neighbor of a predictor cell, the test distance of the centre pixel is set to the current distance value of the neighbor plus seven; and for all horizontal and vertical neighbors, the test distance of the centre pixel is set to the current distance value of the neighbor plus five.

Since the spread of distance from source cell to the neighbor cells in sequence is the nearest distance so the GIS computes all possible path from source cell  Csrc  to  Cx and determines which one has the lowest total distance. This value is found for each cell. To obtain distance values in meters, these raw values are divided by 5 and multiplied by the pixel size and a correction factor.

For 4-connected cell operation, distance between two neighboring pixels along column or row and for 8-connected operation the diagonal distance is also calculated which again is a Pythagorean solution in raster approach.

The ‘basin developed by implementing proximity principles’, as above, can hold good for detailed drainage derivations, but for derivative basin map from less detailed drainage map will not be close to real world outline. Meaning thereby, wider valley will have smaller and a narrower valley will have larger defined watershades. As a corrective measure one need to take an output with buffer zone defined by 30% of the BASIN width and within this

expanded extent of the output, on making a conditional cum neighbourhood scripting for directional value of flow for  highest height (from DEM) among the neighbour cells to the central cell, followed by area numbering for the directional zones developed this way.  A clipped out DEM from the buffered area can be used for the purpose. A statement for alloting a common value for zonal cells of interest generates the BASIN.  This output for basin can be used as (a raster layer or vectorised layer) an operational layer for  raster calculator statements or clip operations, one after another,  result into clip-outs of other criteria themes for carrying out spatial analysis within the spread of the basin.