Unfortunately I am only somewhat GIS literate. I usually would discuss questions like this with GIS expert. But if that not possible, the old school methods are slow, but can work. In GIS, I would go back to original contour lines, and adjust them manually to new field elevation data, then make another DEM and convert to raster. Hopefully, GIS experts have a way to automate adjustments like this.

Another, probably easier but more cost, would be to conduct aerial LiDAR Survey, and use any field ground points to help adjust and rectify LiDAR coverage, convert that to DEM and raster. In a researchgate coauthored paper in 2007, I think figure 4 compares a LiDAR coverage with standard topographic map contours. It is not unusual for standard cartographic topographic maps to have substantial error in areas with dense forest cover, and also flat terrain detail of coastal plain, floodplains and valleys may be lacking.

hello , i think one of the good solution is to resample the raster by using resampling technique like nearest neighbor resampling. The result is only as precise as the coarsest input.

In ArcGIS the following are the two best ways you can do that.

0-Almost similar tools exists in QGIS, Global Mapper and RASMapper (HEC-Ras GUI pre processor), also you can tell Tuflow automatically to consider merge of two raster for calculation (But probably it breaks or give instability in the boundary if they have different cell size).

1- If you have Spatial Analyst license of ArcGIS, your best bet is "Raster Calculator", here you can do numerical calculation cell by cell on two DEM values (for example you can write expression and tell where this is higher get the average where this is lower just use the other one ....)

2- If you do not have Spatial Analyst license in ArcGIS you can still use "Mosaic to New Raster". Here you can tell it which one is your priority in the overlap area. Also you can choose the resolution of the output if they are not in the same resolution (Specify Cell Size box in the picture). You have some functionalities of Raster calculator if you open "Environment bottom" in this toll box (not all thou!)

Hope this helps,

Kaveh

Every DEM is a model, in which elevation data are estimated from sample data. In case of square-cell grid DEMs, the output elevation of each cell is interpolated from the elevation dataset (in which sample data represent input data). The accuracy (the difference between the actual elevation of the surface and the estimated output elevation) of the interpolation process and output model depends on many factors, including the spatial density and distribution of input sample points. The RMSE (Root Mean Square Error) gives a measure of such an accuracy.

In addition to the above practical suggestions, you may verify the accuracy of the available DEM (if not provided by the author of the interpolation) by matching your field collected sample data with elevation data given by the DEM. If you have the original input elevation data of such a DEM, you may add your field data and manage with a new interpolation.

Consider that you are working in a river environment in which topographic changes may be very rapid, so the inaccuracies of the given DEM, compared to your field collected data, may also depend on the time passed between the two surveys (the survey of the elevation data processed to output the given DEM, and your field-collected point survey), specifically the effects of related geomorphological processes that acted in the time interval, including erosional/sedimentary processes connected with the evolution of both braided and meandering rivers.

I think it will be useful,first convert the point data to raster using interpolation technique.Check resolution of the DEM.Then re-sample the interpolated data into the size of DEM.Merge the DEM and Re-sampled data using merge tool.

Thank you all!

@ Suraj Shah Do you know how can i merge aster dems in arcgis? when i use ''mosaic to new raster'' there are black borders and can't merge well like this image that i attach:

@ Suraj Shah