A three-step analysis of the cost-effectiveness of the streamflow-gaging program in Minnesota is documented in this report.

In the first step of the analysis, the data uses and funding sources were identified for the 96 continuous-record streamflow-gaging stations operated in 1985. Nineteen sources of funding and 42 uses were identified for the data collected in this program. Two stations were identified as producing data no longer sufficiently needed to warrant continuing their operation. Three other stations were identified as having uses specific to short-term studies. One station was destroyed in 1985. It is recommended that the remaining 90 station be maintained in the program for the foreseeable future.

In the second step, multiple-linear-regression analysis was investigated as a possible method for providing the data collected at 23 stations. The multiple-linear-regression method was not sufficiently accurate to provide the needed data, and it is recommended that the 23 stations remain in the program. It also is recommended that flow-routing methods be investigated to see if they could provide the needed data for stations on the Red Lake River, the upper Minnesota River, and on the Mississippi River in the Minneapolis-St Paul metropolitan area.

In the third step, the cost-effectiveness of collecting data from 77 of the remaining 90 stations was determined for the open-water period, April 1 through October 30. Data for 13 stations are provided to the U.S. Geological Survey or are collected at fixed intervals and were, therefore, not used in the analysis. The average standard error per station for estimation of the streamflow records is about 24 percent for the statewide network, about 11 percent for the stations operated by the St. Paul field office, about 22 percent for the stations operated by the Grand Rapids field office, and about 37 percent for the stations operated by the Montevideo field office.

The current policy for collecting data from the 77 stations during the open-water period cost $198,000 in 1985. The estimated average standard error per station for the statewide network could be reduced from 24.4 percent to 20.6 percent at the $198,000 budget, if the minimum number of discharge measurements at each station were reduced from five to three during the open-water period and the remaining budget were used to make additional discharge measurements at stations with large standard errors.

It is recommended that, before this data-collection plan is implemented, the effects of the plan on the cost of collecting data be evaluated for (1) possible increased lost record because of the data collection plan, and (2) the possible need for additional trips to visit noncontinuous-record stations. It also is recommended that the data-accuracy needs of the funding agencies be considered before the plan is implemented.