ArcGIS si modelare DEM

Trebuie sa modelez un DEM cu ArcGIS, un soft super tare, dar cam asta trebuie sa fac (din ce am trimis unui coleg cu care am mai lucrat, de prin alte tari). M-am cam impotmolit si, cum nu e nimeni care sa iti dea ajutorul daca nu il ceri, am tot testat, ras-testat iar acum sunt in cautare de optiuni.

Am tot incercat diverse procedee, algoritmi, cu unii nu obtin ceea ce vreau, cu altii dureaza prea mult (cam 6 ore a durat interpolarea unui raster de 3,7 milioane de puncte. Inmultit cu 200 de sectiuni, asta inseamna aproximativ 1200 de ore. Cine dracu are de stat doar pt. conversii, 1200 de ore? Bine, si precizia este extrem de mare, din discutiile cu colegul meu a rezultat ca ei au facut interploari ale rasterelor pe la 5 m, nu 0.5 m. In orice caz, voi mai vedea.

**************

The precision, according to our client requests, should be of 0.5 m (vertical accuracy) in flooded areas and 2.5 m in other areas, could be even 5 m in mountain regions, where water is for certain staying within the edges and there is no flood danger (not even in the past). On my opinion, is it a bit too high precision, but that's they want to do the project in this way, it shall be something new in Romania.

Imagine that I will have around 400 (actually 398) maps like those one (DTM+LiDAR+bathymetry combined into a single map) and I have to do the DEM in order to server as input data for software called "MIKE 21 Flood" (working with ESRI grid).
If you need, I can upload a grid generated using Natural neighbor algorythm, on pixel size of 0.5 m (it took my computer around 6-7 hours), but this is huge, for these 2 maps (around 3.16 gb).
So, the main ideas are:
1. To have the 400 maps full of xyz coordinates into one folder 2. To run a script which will convert all of them into ESRI grid into another folder, based on specific map sheet name 3. To take care that, in each map sheet edge, errors are removed (I am not sure if I explained well on this point, but imagine that making such grid on each map rather that doing the grid on whole 400 map sheets introduce some errors on the edges, even I don't think it's possible with my actual hardware) 4. Input into the hidraulic model in order to extract profiles, to calcule water volumes, and specific calculations (this is another chapter, not handled entirely by me)
Of course, I accept any other ideas, it's not necessary that the workflow should follow these procedures, but the final product should be ESRI grid (I will, though, ask about a alternative input data into MIKE), that's why I need your expert opinion on this and, more important, someone's opinion who was worked before with such things, as you did.
**************