THE HISTORY OF THE HYDROMET SYSTEM
What does "HYDROMET" mean? The term HYDROMET is a contraction for the terms hydrological and meteorological. As used in the Pacific Northwest among water resource agencies, it identifies the concept of automatically sensing and transmitting data in "real-time" from an unattended remote site to a computerized base station where it can be easily accessed and analyzed by water resource managers to make operational decisions.
Credit for the original HYDROMET concept can be given to the early day managers of the Columbia River system. It was in the 1960's that a group of Federal, State, and private water resource managers, known as the Columbia River Water Management Group (CRWMG), began to realize a need for better and more timely data to operate the system. It was this group that saw the value of the Columbia River systems water and the demands that would ultimately be placed on it in the future. They agreed to develop a rapid and continuous system for collecting hydrometeorological data that could be easily distributed to all users in the northwest for input to forecasting and operational computer models. This system became known as the Columbia River Operational Hydromet Management System (CROHMS). As envisioned by these early managers, the system would be used for flood control, irrigation, power scheduling, fish and wildlife, and recreation; however, little or no mention was made of a possible use for drought management.
The original communication system used for CROHMS was the Columbia Basin Teletype (CBTT). The CBTT had nodes at all project and operating offices of interested agencies. Each office had the ability to input manually collected data to the system and pull off data from other locations as needed. Some data was transmitted by telephone and mail.
In September 1969, the Regional Director signed the Hyrdromet Data Management Memorandum
of Understanding. This memorandum states that a "real-time" hydrometeorological network
is necessary in the Columbia River Basin in order to realize maximum benefits from
multipurpose Federal water projects. In addition, it states that in order to fulfill our
water management responsibilities, the system should be carefully coordinated among all
interested agencies and data should be channeled to a central location for processing,
dissemination, and storage. The memorandum specifically laid down the following
responsibilities for the agencies to work toward:
(1) To prevent unnecessary duplication of facilities
(2) Provide maximum utility and uniformity in collection and use of HYDROMET data
(3) Ensure the highest degree of reliability at minimum cost
(4) Gain the economics of multiple use existing and proposed facilities consistent with individual agency responsibilities
The memorandum was also signed by Corps of Engineers, Bonneville Power Administration, Federal Water Quality Administration (currently EPA), Geological Survey, Forest Service and the Environmental Science Services Administration.
To insure the development of this network was carried out according to the memorandum, the Columbia River Water Management Group formed a subcommittee called the Hydromet Data Committee, which has representatives from the Corps of Engineers, Bureau of Reclamation, Bonneville Power Administration, National Weather Service, and Geological Survey. Other state and federal agencies attend as needed.
Once again the need for the "real-time" system was justified on flood control, navigation, power, irrigation, recreation, and data record keeping, but even those farsighted managers never made any reference to drought and how the system could be used during such a period.
The memorandum also identified areas of agency responsibility for coordinating the development of HYDROMET networks in certain geographic areas in the northwest. Along with coordination, these agencies were to take the lead in developing the networks and insuring their compatibility with the overall CROHMS system.
Reclamation was given responsibility for coordinating the development in the following areas:
(1) Snake River Basin above Hells Canyon, Idaho
(2) Yakima River Basin, Washington
(3) Upper Deschutes River Basin above Tygh Valley, Oregon
The Corps of Engineers, Bonneville Power Administration, Geological Survey, and National Weather Service accepted the responsibility for coordinating the development of the system in the remaining areas of interest in the Northwest. In addition, the Corps assumed the responsibility of furnishing and managing the CROHMS Central Computer Facility (CCF), which was located at the Corps headquarters in Portland, Oregon. The CCF served as the CROHMS communication interface and central data base for outlying HYDROMET network controllers and the CBTT, all of which furnished data to form the main database in CROHMS.
The HYDROMET network, when completed, was envisioned to include a hydrometeorological station at each dam, at key inflow, outflow, and other river control points and at remote medium and high elevation weather sites. These stations would report to small computerized network controls in project areas which in turn would report their data to the CROHMS CCF. There the data would be available for others to access.
DEVELOPMENT OF RECLAMATION'S HYDROMET NETWORKS
The Yakima Project Office took the initiative to build the first of Reclamation's HYDROMET systems in 1976. The initial system consisted of 16 remote stations including reservoirs, river gauges, canal gauges, and weather stations that reported via a line of sight radio system. The radio system's backbone, which consisted of three repeaters, had been used for the project's voice communication system for years, therefore, it was easy for the HYDROMET system to piggy back the existing radio network. The voice communication system had the added advantage of allowing maintenance staff in the field to talk directly to the operators in the office, when troubleshooting remote equipment.
The network controller was located in the Yakima Project Office and consisted of a Digital Equipment Corporation PDP 11/03 with a customized radio interface for the HYDROMET system.
Acquiring data from remote stations is accomplished in two ways. First, the controller can be programmed to interrogate groups of stations on a regular predetermined schedule, say every 2 hours. This provided for a complete data base to be constructed. Secondly, during emergencies or other periods of concern, the controller can make ad hoc pollings of certain stations in question as often as the operator requests. If at any time a report gets interfered with, the operator can just repoll the station.
In 1979, Reclamation procured the Boise-Minidoka HYDROMET System for the middle and upper Snake River Basins. Although originally planned to be two separate systems with a network controller at both the Minidoka Project Office in Burley, Idaho, and Central Snake Projects Office in Boise, Idaho, the two projects agreed to combine the two systems into one, thereby saving considerable cost. The system was originally envisioned to be some type of land-based radio system, similar to Yakima's, however, the proposals received from the contracting community included systems supported by VHF radio, microwave, satellite, meteor burst, and combinations of two of the above.
Evaluations of the original proposals by the Technical Proposal Evaluation Committee (TPEC) and following through normal contracting procedures to the request for Best and Final Offers, it was determined that a Geostationary Operational Environmental Satellite (GOES) based system proposed by Sutron Corporation had both the best technical score and the lowest price. In August 1979, an award was made to Sutron for the Boise-Minidoka HYDROMET System. The original contract included 67 remote stations consisting of satellite radios, electronics, and sensors of various types; a 7-meter Direct Readout Ground Station (DRGS) and associated electronics in Boise for the GOES; and a CCF consisting of a Digital Equipment Corporation VAX 11/780 to serve as the network controller for communications and data processing. The system was provided as a turnkey system with Sutron responsible for furnishing and installing all equipment and software.
The satellite system had several advantages over a land-based radio systems in this particular project area. First, it allowed easy coverage of this large mountainous geographic area without the need to maintain a large, expensive and complicated land-based repeater system. Secondly, it allowed unlimited expansion throughout the entire northwest for future HYDROMET development at no extra cost for the backbone communication system. Thirdly, it provided for event or random reporting of significant events as they occurred without prior notice. This will be discussed in more detail later.
There are also some disadvantages or tradeoffs to be considered when compared to land-based radio systems. First, stations can not be interrogated since the communication system is one-way from the remote station to the DRGS. Secondly, voice communication is not available for field maintenance and troubleshooting.
Until 1979, the GOES operation was strictly a time ordered system that basically contained only self timed Data Collection Platforms (DCP). The only exception being a few interrogatable DCP's that were expensive and difficult to maintain. They could only be interrogated by National Oceanic and Atmospheric Administration (NOAA), at the request of the users, through the main NOAA ground station in Wallops Island, Virginia.
As mentioned earlier, the system selected by Reclamation is unique in that the DCP at each remote site is microprocessor controlled and has the capability to transmit through two channels on the GOES system. One channel handles only "self-timed" transmissions, whereas the second channel is dedicated to "adaptive random" transmissions.
Operation in the self-timed mode is as follows. The DCP interrogates all sensor outputs at some user defined interval, say every 15 minutes, and stores the values in its memory. At a preassigned time interval every 4 hours, the DCP transmits all stored values (16 values from each sensor for a 15-minute update) to the CCF through the DRGS in Boise. This produces a very complete detailed data base for record keeping and publication purposes. It is also possible to send redundant data so that each transmission contains the latest 4 hours of data and the previous 4 hours for a total of 8 hours of data.
Transmissions in the adaptive random reporting mode are completely unscheduled with the decision to transmit being made by the DCP. This is accomplished by programming threshold values (e.g., flood stage) in the microprocessor which the DCP uses to compare with current sensor outputs. If the threshold values are exceeded, the DCP computes a random transmission rate and begins to transmit randomly. The microprocessor also computes rates of change between sensor readings, if the rate of change exceeds the preprogrammed threshold values (e.g., river state rising at 0.2 feet per hour), this also causes the DCP to compute a random transmission rate and begin transmitting.
Each time the DCP transmits randomly, it only send three values; the most current value and the two preceding values. Also, when the DCP goes into random mode it will send at least three random transmissions before shutting down. However, if the threshold values are continually exceeded and/or the rates of change increase, the DCP will continue in the random mode until the situation returns to normal. It is important to note that as the rate of change of the sensor value increases or a second level threshold value is exceeded, the random transmission interval is shortened thereby transmitting more data as the event becomes more serious. The DCP can be used to monitor critical low flow conditions by reversing the process and using low threshold values.
All data received by the CCF is immediately processed and stored in a file called DAYFILES where it is available to users through time-share terminals. At 0500 hours each morning, the CCF compiles data from the previous days DAYFILES data to be stored in the ARCHIVES file. The ARCHIVES data base includes midnight reservoir elevation and contents, maximum and minimum temperatures, and mean daily flows, etc. This archived data is then also available to users through time-share terminals.
USE OF HYDROMET TO MANAGE AND ALLEVIATE DROUGHT EFFECTS
As pointed out earlier, drought management was not an identified benefit of developing HYDROMET networks in the northwest by the early day managers. In fact, it was just the opposite, flood control was recognized as a primary function of the real-time telemetry systems they envisioned. For instance, funding for Reclamation HYDROMET systems was 100-percent nonreimbursable flood control monies, all other benefits to the system got a free ride. Funding with flood control funds was easily justified by projected damages that could be alleviated with the real-time networks. This was validated in the spring of 1981 just as the Boise-Minidoka system was being completed. The spring runoff was just about over and most reservoirs in the upper Snake were filled when a 2-week period of heavy rains fell in the area. By carefully watching inflows to the reservoirs, the managers were able to increase outflows slowly, performing a delicate balancing act. After the flood had passed, the Minidoka Project Office estimated that the HYDROMET system had been the primary reason they had been able to control flows through the City of Idaho Falls, where a powerplant was under construction, thereby, preventing approximately $1 million in damages.
Since storage facilities are one of the primary aspects of drought relief programs, and since 80 multipurpose storage reservoirs were already built in the northwest, it makes sense that drought was not considered as a primary benefit of the HYDROMET networks. Also, serious water supply shortages had not occurred on a widespread basis in the northwest since the completion of the major storage projects were built during the period from 1900 through 1940;s. Although the northwest has encountered some below normal runoff years since the 1940;s, water shortages to users have been minimized by the previous years carryover in the storage system. It wasn't until now when water years 1987 and 1988 were consecutively below normal that the storage system was stressed.
Water year 1987 and particularly 1988 have seen the HYDROMET system used to manage the available water supplies to their maximum potential. Following are some of the applications Reclamation has used to help alleviate the drought's affects:
It is obvious during a drought or any period of water shortage the key is to use the supply thats available to its maximum potential and avoid waste. The real-time HYDROMET systems are an excellent tool to help accomplish this difficulty task. With river gauging stations reporting every few hours, it is easy to keep track of each cubic foot of water in the system. The more stations the more finite the control can be. In Reclamation's Boise-Minidoka HYDROMET system, automated stations are located on each major inflow to the reservoirs, on the reservoir itself, downstream to measure river discharge, and quite often major canals that divert from the reservoir directly, and finally at critical points on the river system throughout the irrigated areas.
The advantage of such control was realized by the Idaho State Watermaster for District No. 1 in the Upper Snake River Basin in 1983. At that time, he requested permission to add major canal diversions to the HYDROMET system. The first year five major canals were installed on a pilot project basis. The program proved to be very successful and has expanded to include 33 stations to date with plans to expand to around 60 in the future. Presentation of the benefits of real-time monitoring for water management was presented in the paper "Data Automation for Water Supply Management" by Sutter, Carlson, and Lute at the 1982 ASCE Conference in Lincoln, Nebraska.
If you add a series of real-time meteorological stations to the benefits already discussed to be derived from a hydrologic system, you achieve a second level of drought management benefits. Now you have the advantage of knowing where and when precipitation is falling and can anticipate areas that may cut back on irrigation water supply requirements. This can allow the managers to retain water in the reservoirs that would normally have been wasted downstream. In most irrigation operations, there is a downstream control point that a manager wants to control to zero or some required minimum flow past that point.
A third benefit to drought management is the ability to run more frequent forecasts of the expected water supply. Under the old manual data collection system, forecasts were only run on a monthly basis since that was the only time data was available. With the advent of real-time data, forecasts can be run daily, weekly, biweekly, or monthly. This helps to give the managers a better handle on what is going to be available and can help farmers plan the quantity and type of crops to plant. As an example, in 1988 based on a forecasted low water supply, some farmers chose not to plant all their land, thereby concentrating their total water supply on a reduced acreage. Having the availability of the real-time data also allows the managers to more accurately calibrate forecast models for mid-month runs.
One of the newest uses of real-time data for drought relief is in the field of computerized irrigation scheduling.