THE USE OF SRTM DIGITAL TERRAIN MODEL FOR

THE USE OF SRTM DIGITAL TERRAIN MODEL FOR THE SPATIAL REPRESENTATION OF LIQUID AND SOLID RUNOFF – ALTITUDE LINKS Viorel CHENDEŞ NATIONAL INSTITUTE OF HYDROLOGY AND WATER MANAGEMENT Bucharest, ROMANIA

GENERAL INFORMTION CONCERNING SRTM n n n NASA Shuttle Radar Topographic Mission (SRTM) is a joint international project. The original data have a resolution of 1’’ (aproximately 30 m). The free online data have a resolution of about 90 m (3’’). The SRTM model was developed through the use of “radar interferometry” technique.

DTM FOR ROMANIA AT A MEDIUM RESOLUTION n n n Being based on RADAR images, the SRTM data include "no-data" spaces where the water or the shadows obstructed the determination of the altitude. Most of them are spread along Danube, Danube Delta, the lake areas and in large floodplains. They are restrictive factors especially in hydrological modeling applications. The correction of the SRTM model was done in Arc. View, using mostly the Spatial Analyst extension.

SRTM: 90 m 30 m n n n The initial resolution was approximately 90 m. For using at detailed scales and in different applications, the SRTM model was re-interpolated and exported with a 30 m resolution. Even though the elevation errors for each pixel can not be eliminated, being the same at 30 m as at 90 m, there are several benefices. Mainly they consist in a clear delineation of the relief forms (valleys, slopes etc. ), fact that helps to reach better results in hydrological applications.

SRTM at 3’’ res. (about 90 m) Hm = 1275. 3 SRTM at 30 m from 3’’ SRTM Hm = 1279. 7 DTM from topographic maps at 1: 25. 000 scale Hm = 1267. 3

SPATIAL REPRESENTATION OF MEAN SPECIFIC LIQUID RUNOFF – ALTITUDE LINKS n n q (l/s*km 2) eliminate the influence of the surface. q is a very useful parameter in spatial analyses, indicating the physical-geographical conditions and their influence on the discharge. The Southern Carpathians, relief unit chosen for the exemplification.

Step 1. Completing the data base n Generally, in Romania, the hydrometric stations offer data for a long period (1950 -2005) and only in few cases there is needed an extrapolation of the data. Correlation of the mean annual liquid discharge (m 3/s) between two neighboring stations with similar physical-geographical conditions Correlation between the mean annual liquid discharge (m 3/s) registered over 1975 -2000 period with the one registered over 1950 -2000 period

Step 2. Regionalization of mean specific liquid runoff n The regionalisation of the runoff using the correlations q-Hb is carried out using classical methods. In the Southern Carpathians, even if it seems a homogeneous area at a first view, we could identify 3 correlation functions q - Hb: Zone I: 0. 0339*H 0. 96 Zone II: 0. 0245*H 0. 96 Zone III: 0. 0177*H 0. 96

Step 3. Mapping the q - Hb correlations zones n This means the map out of the zones for every correlation curve in accordance with the stations which define them.

Step 4. Building the mean specific liquid runoff map n n n This map is built on the basis of SRTM model and of the q-Hb functions For this we used ESRI software products (Arc. View 3. x, Arc. GIS) which allow application of a function in which a variable is a grid (in our case - the altitude). These tools are offered by Spatial Analyst extension, (Map Calculator in Arc. View 3. x or Raster Calculator in Arc. GIS).

Step 4. n For areas characterized by a single correlation function the building mode of the map is simple, by applying once the command in Raster Calculator.

Step 4. n In the case of more correlations, there are two ways of building this map: a) step by step a 1) applying in RC one of the q-Hb functions, using the DTM of entire area as input. It will result an intermediary grid qn(the mean specific liquid runoff grid), where n is the zone number. For example, for zone I is applied next syntax: n n (0. 0339*[Elevation]. Pow( 0. 96)) (in Arc. View 3. x) 0. 0339 * Pow([Elevation], 0. 96) (in Arc. GIS)

Step 4. a) step by step a 2) The converting of the polygon which represents the limit of the zone in grid using an attribute by 1 value for that zone and by 0 value for the rest of the zones Zone I (value “ 1”) Other zone (value “ 0”)

Step 4. a) step by step a 3) The multiplying of the two rasters. It is obtained a final qn raster, where n is the zone number. This will have the value of the mean specific liquid runoff for the inside pixels from zone n and 0 for the rest of the pixels x = a 4) Repeat a 1 -a 3 steps for the others zones. For each zones is obtained a final qn raster. a 5) The adding together of the n grids by former type, resulting a final grid which represents mean specific liquid runoff

Step 4. b) Starting from the previous way, the steps were grouped in a single expression which can be used in Map Calculator. n The needed layers are: - [Elevation]: SRTM DTM - [Zone_q]: raster conversion from the polygon theme which represents the q - Hb correlation zones. This raster will receive values equal to the number of zones, all the pixels in zone ‚n’ having that ‚n’ value. n relational operators (=, <, <=, <>, >, and >=) evaluate specific relational conditions. If the condition is TRUE, the output is assigned 1; if the condition is FALSE, the output is assigned 0. So, the syntax [Zone_q] = 2. As. Grid or [Zone_q] = 2 (in Arc. View 3. x) will create a grid of value ‘ 1’ (for the pixels inside zone 2) and ‘ 0’ (for the pixels outside zone 2).

Step 4. n In this way, the syntax for the computation of the mean specific liquid runoff in the area of the Southern Carpathians, based on the following correlation functions q - Hb Zone I: 0. 0339*H 0. 96 Zone II: 0. 0245*H 0. 96 Zone III: 0. 0177*H 0. 96 might be written as follows: ([Zone_q] = 1. As. Grid * 0. 0339. As. Grid * [Elevation]. Pow(0. 96)) + ([Zone_q] = 2. As. Grid * 0. 0245. As. Grid * [Elevation]. Pow(0. 96)) + ([Zone_q] = 3. As. Grid * 0. 0177. As. Grid * [Elevation]. Pow(0. 96)) or, simpler ([Zone_q] = 1* 0. 0339 * [Elevation]. Pow(0. 96)) + ([Zone_q] = 2 * 0. 0245 * [Elevation]. Pow(0. 96)) + ([Zone_q] = 3 * 0. 0177 * [Elevation]. Pow(0. 96))

Step 4.

CONCLUDING REMARKS There are still problems to be solved, the principal one being the abrupt pass from one correlation zone to another. It is necessary to find a method to delineate some. THANK buffer zones YOUthrough which the pass should be smoother. FOR YOUR ATTENTION n Another point refer to the numerous methods of processing and analyzing the map: n the analysis of the mean runoff distribution by major relief units or by altitude steps, n the analysis of runoff formation, n the amount of the water volumes, n the distribution of the water volumes by altitude steps, n the cumulative values of the water volumes, n These aspects are to be presented in a future paper.