02-05-2019 08:06 AM
You have not 'optimal' setting for this. You can consider all the cells of your Study Area as part of the hydrological network... The only goal of this Recode operation is to keep the bigger values, and the threshold depends of your local network.
Perhaps you can get more 'logic' results by using the hotspot detection command and not the recode one. The hotspot detection allows you to define a threshold which is proportional to the average value of all the cells