Louadj,

You need to have your data in XYZ format where X = easting coordinate, Y = northing coordinate, and Z = salt concentration value. These data then need to be brought into a program like Surfer or Voxler in order to create a DEM. The DEM then can be overlain as a color distribution map on top of topographic data such as a topo map with the colors representing salt concentration value ranges.

Dear Louadj Yacine,

The answer is very much depending on WHAT you want to study. A Digital Elevation Model (note the spelling) is in general a model of a surface (a real one, but it can also be hypothetical) and reflects the elevation (with reference to a certain base level) of the surface through the model space. Here, the "model space" stands for the area you want to model/study.

If I understand your project, it is about to model certain soils. This is important, because you need appropriate data for the area with the requested resolution and accuracy.

To be able to answer your question the following has to be decided:

1. How large is the study area? (And where it is? Are the multiple locations/sites?)

2. What is your needed resolution? (Practically the smallest distance that you want still recognize in your model, or the resolution of your DEM grid.)

3. What is your requested (vertical) accuracy?

And in relationship to all above:

What is the data acquisition technique that can provide the data fulfilling the aforementioned criteria and can be produced in time and within you cost budget.

(Or: Do you have already the measurements?)

Without knowing the details of your project, for studying the soil surface in general, good resolution (1 cm or better) stereophotogrammetric image sequence (taken e.g., from a kite or from an elevated mobile plateau) or laser scanning (a.k.a. LiDAR) data (terrestrial laser scanning (TLS) or low-flying ULS (= unmanned laser scanning, i.e drone-mounted laser scanning) could be suitable.

Kind regards, Balázs

thank you  Dr Székely and Dr green for your answer,  For the measurement , i have some result  concerning the EC ( soil : water ratio 1:5) of 8 soil profile at three depth 0-20, 20-40; 40-60 along a toposequence ( North-South) , can i use this profil with their location ( X, Y, Z) to construct my DEM or i need more profile ?, i should add the Area of my study cover about 352 square Km.

Louadj,

Normally to construct a DEM or contour map of an area you need a sufficient number of XYZ points, otherwise the spatial representation might not be within the realm of actuality. These points also have to bound and infiltrate an area, thus a straight line of data can not be used.

Before undertaking an arduous task of collecting more data, my suggestion would be to generate a DEM or contour map of you data, then select a few random locations to ground-proof the model. This will show you if the number of data points that you have is sufficient or whether more points are needed.

It seems that you need an individual DEM or contour map for each of the depths in order to display your data. In that case you would likely need to stack the individual DEMs or contour maps over your landform map in order to display all the data at once, or you can show each depth individually.

Another option, given the few data points that you have, is to show each profile location on a topo, along with a vertical bar graph showing the corresponding data values at different depths.

Louadj,

The number of necessary points also is dependent on slope and complexity of landforms. If you are looking at the depth of salt saturation of soils in an area after a tsunami or typhoon storm surge then that assumes a flat or slight slope from shoreline to maximum inland extent of saltwater intrusion. In that case, then the number of data points can be far less than, say, measuring the natural salinity of soils in a upland forest setting.

Mind if I ask someone with the Soil Service for input?

Dr Green, 

How can i generate a DEM of my study area with few data that i collect ( 8 soil profile with 3 depth that makes 24 Samples) ,i use the software that you mention in your first answer ( Surfer) or Voxler?

For either program you take your data and put it in XYZ format. Then separate the data by depth, saving each so that you have all of your XY 0-20, XY 20-40, and XY 40-60 in different tables. Bring each of these into Surfer one at a time and have the program grid the data. You can then have it display that data as contours, 3-D representation, DEM, etc. These can be saved as georeferenced shape files or TIFs. For the DEM or 3-D models you can play with the color palette or design your own in order to get the desired effect.

Voxler is a lot more complicated and really requires you to read the instructions. However, it is capable of rendering data in 4-D. In this case 4-D means XY coordinate, Z depth, and Q quantity. In this way you can show all of the data within a 3-D matrix. But, I caution that it is difficult to achieve the desired results if you have never used Voxler. You can also do the same things in Voxler that you can in Surfer.

In addition, the Surfer program should be Version 11 or above. Earlier versions required the data points to bound an east-west or north-south square or rectangle. The program would render data from irregular or non E-W N-S areas into square or rectangular plots, thus distorting the data.