Dear Damepaia S M Pdah , can you give us a little bit more information about the spatial location of your research area? DEM gives you only surface information. Therefore, maybe you should think about sub-surface networks as the karst systems.

DEM itself does not affect drainage pattern. It's the topography, climate condition, soil properties and land use conditions that play the critical roles in affecting water cycle. However, if you mean how DEM would affect the DEM derived drainage system, I would say it's the DEM resolution. Your delineation will be more accurate with higher DEM resolution.

In my researchgate, there are a few papers I and associates worked on, that might be interesting to you. Check out Stream Types and Management Implications which used and compared stream densities based on 1:24000 and 1:100000 scale based DEMs. A couple more recent papers (James ET al, and Maceyka and Hansen) used LiDAR based DEMs with much greater detail for gully, channel and watershed boundaries. LiDAR in one of my projects in Georgia was able to detect intermittent or ephemeral channel detail of an added two to three orders, as compared to 1:24000 topographic map based DEM. LiDAR based DEM in heavy forest also detected old drainage ditches, areas of past channel relocation for farming, old skid roads and other features. LiDAR gives a much truer georeferencing of stream network location and even roads on topographic maps. Field work and experience in area of interest helps to identify the likelihood of flow permanence, recognizing there can be substantial variation in size of headwater catchments that produce ephemeral, intermittent or perennial flow.

Concerning your our basic question, I am probably not the expert on how a DEM works, but basically the conversion of a topographic map into a DEM, divides the land into small pixels, such as 30m x 30m or 10m x 10m. Each pixel is part of a detailed raster network, with its own elevation, location, etc. The hydro models are designed to help find the flow path as the water accumulates as it flows downhill seeking the lowest direction as it moves downhill. Expect to find some mistakes if not using LiDAR based DEM. Cartographers are not perfect in photo interpretation, especially in heavily forested areas. LiDAR can also be helpful as water absorbs the laser beams, useful for identifying surface water at time of survey, such as streams, wetlands, etc.

As mentioned, subsurface flow is another subject that can be important in some situations As karst geology, but LiDAR or aerial photo DEMs might give some indicators as sinkhole terrain. Experienced geologists, soil scientists, ecologists and hydrologists can sometimes read photos and topographic indicators with greater intensity and interpretative meaning.

The Digital Elevation Model (DEM) alone will not give information about a water cycle. It is generally used by a Geographic Information System (GIS) program along with other spatial information. A good hydrology model running in a GIS is needed to use that DEM, along with soil and vegetation layers, and climate information in order to estimate the various components of the hydrologic cycle.

It requires basic geographical information of that area where you want to conduct survey and also you need to compare with ground data by consideing NDVI and other values LULC

The accuracy of drainage network depends on DEM resolution and pre processing (fill, sink etc). The overall accuracy of srtm v3 is good.

Land use, land cover, wind direction and speed, other climatic variables, terrain and flows effects

Yes, I agree with Dr.Aslam