HelioClim and the Elevation Databases: Terrain Base 5' (TB5) and SRTM

A problem of units ?
of terminology ?
Consult the Education service.

TB5 and SRTM in the solar radiation framework
Free Access to the SRTM database
Effect of the elevation database on maps
For Mines ParisTech and Transvalor only

TB5 and SRTM in the solar radiation framework

Why are topographic data important in the solar radiation framework?
These data can be used to modulate the irradiation values of the HelioClim databases. Two main origins can be identified: the variation of the optical path length and the shadowing effect due to the horizon. This page deals with the first of these two origins. The principle is the following: the bigger the height, the shorter the optical path length, and finally the higher the radiation value.

HelioClim and the elevation databases: the Global Horizontal Irradiation values stored into the HelioClim databases (HC1 and HC3) are by default modulated with the heights of TB5. It was a choice made in the past, when the capacity of storage of MINES ParisTech was unsufficient to welcome the whole SRTM database. Our old database HC1has this default height and with no opportunity to modify this characteristic. The enhancement of HC3 compared to HC1 is that a layer of computation has been added to takes into account the higher resolution database SRTM. When a user makes his request and set the elevation to its default value of "-999" (for "use default digital terrain elevation database value"), the user retrieve the elevation value of SRTM.

Illustrations: the second (respectively third) column depicts the TB5 (resp. SRTM) values for each MSG pixel over Corsica. In the case of TB5, several MSG have the same TB5 value since its resolution is coarser than MSG, contrary to SRTM.

Ex. Corsica	TB5	SRTM
Elevation values in the MSG projection
Spatial coverage	Worldwide: [-90°S;+90°N] [-180°W;+180°E]	(idem)
Spatial resolution	5’ = 5 minutes = 10 km	Approx. 100 m

Top of page

Free Access to the SRTM database

You have the possibility to download for free and by yourself SRTM elevation maps over any area worldwide. The SRTM image in the WGS 84 projection for Corsica on the left has been obtained using the following webservice (Firefox, usually saved on the desktop):
http://www.webservice-energy.org/mapserv/srtm?SERVICE=WMS&VERSION=1.3.0 ... &REQUEST=GetMap&BBOX=41.2,8.3,43.2,9.8&CRS=EPSG:4326&WIDTH=600&HEIGHT=1000 ... &LAYERS=srtm_s0&FORMAT=image/tiff

http://www.webservice-energy.org/mapserv/srtm?SERVICE=WMS&VERSION=1.3.0 ... &REQUEST=GetMap&BBOX=[min_lat,min_lon,max_lat,max_lon]&CRS=EPSG:4326 ... &WIDTH=[nb_pix]&HEIGHT=[nb_pix]&LAYERS=[srtm_version]&FORMAT=image/tiff
  with, min_lat: west bound of your bbox
  max_lat: east bound of your bbox
  min_lon: south bound of your bbox
  max_lon: north bound of your bbox
  nb_pix: size of the image in pixels
  srtm_version: srtm_s0 (full resolution, 16 bits), srtm_s2 (½ resolution srtm_s0), srtm_s4 (½ resolution srtm_s2), or srmt_s8 (8 bits, ½ res. srtm_s4)

Characteristics of the webservice: the webservice enables you to retrieve a map of max 2048 by 2048 pixels. Where SRTM is not defined, i.e. in the water areas, the value is -32768 or -2ˆ15. The highest spatial resolution is 5°/6000 pixels, which means that 1 pixel corresponds to 0.000833° (= 3" = 3*1/3600). If you request a pixel size higher than the spatial resolution, an undersampling operator is applied to create a coarser image. Whereas, if the requested pixel size is smaller than the spatial resolution, a bilinear oversampling is applied on the image.

Please note that other resources/webservices are also available on the site webservice-energy.org. Feel free to explore the potentiality of this link. We remain available to provide you the corresponding support.

NB: There is a bug with firefox which returns a QuickTime error at the end of the process, without the possibility to "save as" the output tif image. The solution is to do that with the wget command:

wget -O image.tif "http://www.webservice-energy.org/..."

Full resolution of SRTM over Corsica

click here to magnify

Top of page

Effect of the elevation database on maps

	a_ Yearly average irradiation map of GHI over Togo with TB5 as default height	b_ TB5
Since October 2009, we provide a service of map creation to our Customers. This service is on request only. One of our Customers asked for a map of annual average Global Irradiation on the Horizontal plane over Togo. We thus created the yearly average GHI map as depicted in the illustration a. We were quite upset while observing this map and notice that larges squares of tint areas spread along a large line oriented along the southwest-northeast direction. After a reflexion with the team of research of MINES ParisTech, the origin of the problem was the coarse resolution elevation database (Terrain Base ´5´) used by default to create the HelioClim database. Actually when we observe the TB5 image over Togo in the illustration b, the same squarred artifacts are visible at the same location of the radiation image, which confirmed the first assumption. We thus ask one of our technical experts to rebuild the HC3 database for the service of map creation with the elevation database of high spatial resolution named SRTM. Indeed, the illustration d which depicts the SRTM image over Togo does not show anymore the artifacts anymore, as well as the corresponding image of radiation using this SRTM default elevation database in illustration c.	(click on the image to magnify)



	c_ Yearly average irradiation map of GHI over Togo with SRTM as default height	d_ SRTM

Top of page

For Mines ParisTech and Transvalor only

In the purpose to correct the mask of sea (255 inland, and 0 outside), we realized that the swath of the SRTM layer used to check this mask was not in agreement with the HelioClim one.
In our pupose to build a correct SRTM layer, we realized that:

The MSG images (named yyyymmddHHMM-msg-ch01.png or -ch02.png), TB5_MSG.bin and the old SRTM image had their first non nul pixel on the western direct (left on the image) at the 56th pixel, 54th on the north, 3658th on the east (55th from the right edge) and 3660 on the south (53 from the bottom edge).
Id_to_Mat.mat, its equivalent in C tab_Id, the HelioClim images db_irr_1305101200.tif, the TB5 from tab_Id, the rhog108.bin (rhog[_day_in_year].bin) are: 86th, 92th, 86th, and 92th. Please note that the Kcmax also has the same swath except that you might have Nan values on the north

The routine to create this correct SRTM layers is named create_SRTM_layer.m. It was posted in svn sometime around May 22nd 2013.

Top of page