A defining feature of dryland crop production in the inland Pacific Northwest is annual, multiple freeze thaw events occasionally associated with snowmelt, rainfall, or both.  These events can produce substantial runoff and soil loss.  Measuring overland flow and soil erosion from these events is a challenge because field locations are not located near power sources,  events are infrequent, event magnitudes are highly variable, equipment freezes or is otherwise impaired due to low temperatures, and sediment loads can be high and impair equipment function.  Since 1997, we have measured runoff and soil erosion from 0.05 ha fields on slopes from 2 to 6 percent, a 1.43 ha hillslope with a 30 percent slope, and small drainages up to 25.2 ha with internal slopes up to 30% and a relief ratio 0.08.  We determined flume or weir sizes using the SCS curve number method with a100 yr rainfall return period.  Despite an appearance of over-sizing, one event justified this decision. We began making flow measurements with H-flumes, transitioned to Parshall flumes and finally settled on the drop-box weir as our instrument of choice.  The main advantages of weirs were its ability to deal with high sediment loads and remain relatively ice free.  Our ability to record stage remained somewhat problematic, and providing sufficient fall below the weir required limited excavation. We collected suspended sediment samples using storm water samplers and after several iterations arrived at a satisfactory method for initiating sampling by using a liquid level switch installed in a special approach section we designed for the drop box weirs.  Collecting grab samples was necessary to provide calibration to stage recordings that were subject to temperature dependent digital signal drift. Ideally, weirs, sediment samplers, stage recorders, and data loggers should be housed in heated structures.