Cold Regions Hydrology Workshop
Breakout Session Description and Status
Threshold Criteria for Ice Jam Potential
During the NWS 2002 Cold Regions Hydrology Workshop (CRHW) there were several presentations and discussions about ice jams and how the NWS could improve their ability to forecast them. During one of the break-out sessions a cooperative project between the U. S. Army Corps of Engineers Engineer Research and Development Center Cold Regions Research and Engineering Laboratory (ERDC/CRREL) and OHD was formulated to achieve this goal. The project had two main components: First was the development of a general threshold algorithm for ice jam occurrence that could be applied across the northern US and Alaska. The second component was a modification to FLDWAV so it could simulate the effects of an ice cover on river stage and flow.
Funding was allocated by OHD for these two projects and the preliminary results were presented at the 2004 CRHW. These presentations can be seen on the CRHW website.
The results from these studies were very encouraging. CRREL related accumulated freezing degree days (AFDD) and a normalized rate of streamflow rise to the occurrence of ice jams at 12 locations throughout the country. General patterns emerged suggesting that this type of approach would enable forecasters to categorize the ice jam potential as low, medium, or high. OHD verified its FLDWAV modifications for a simple ice cover case; however, more complex ice situations need to be addressed.
Representatives from several RFC's and WFO's met during the breakout session to discuss the next phase for these projects. There was broad agreement that the NWS should start modeling ice jam potential using the algorithms supplied by CRREL. Such a model would greatly enhance the NWS's ability to forecast the occurrence of ice jams. Furthermore, it is possible that some of the time series information generated in the ice jam forecasting code may be useful in the future as input to FLDWAV/SHRT. For example, an estimate of ice thickness and ice formation/break-up timing could be used to characterize the ice cover in a dynamic routing model.
It made sense to the group that the modeling should occur within the NWS River Forecasting System (NWSRFS). NWSRFS contains many of the time series needed for the analysis and the framework necessary to calibrate the model and run it operationally. However, calculation of AFDD and the normalized rate of streamflow rise requires new operations within NWSRFS. In order to give the new operations broad application, we thought it would be best to make them generic in nature, so they could be used for other purposes as well, such as reservoir routing or diversions.
The following bullets describe the algorithms for computing the variables described above.
Accumulated Freezing Degree Days (AFDD):
1. Each time period compute the difference between the air temperature and 0°C. (0°C– Ta).
2. Add this value to the AFDD from the previous period to get the AFDD for the current period.
- The AFDD cannot be < 0°C.
- Carryover for AFDD will be required
Flow During Freeze-up (Qf):
Qf = Mean streamflow during the period when AFDD is less than AFDDf , where AFDDf is a threshold AFDD value identifying the freeze-up period.
OR
Qf = Average of the lowest x days of streamflow during the freeze-up period.
Normalized Rate of Streamflow Rise
dQ/dt = (Qi - Qf)/(Qavg ∙ Δt)
Where Qi = current flow; Qavg = long term mean daily flow; Δt = duration of the rise.
To compute these variables within NWSRFS we would need the following capabilities:
Develop a general time series calculator to compute quantities supporting ice jam forecasting.
- Must work in both calibration and forecast modes.
- Compute time-averaged or cumulative quantities. (This will require carryover.)
- Compute differences between time series, variables, and/or constants.
- Need ability to define constants
- Need ability to construct logic based decision trees.
- Need ability to modify time series during run-time.
- Need a function that will look a specified number of periods back in time to determine an average rate of change in a time series.
- Need to compute a square root, so we can estimate ice thickness.
Action Items
Brian Connelly will provide a list of requirements for a time series calculator to Bob Cox (MBRFC) to aid in the drafting of a requirement for a generic time series calculator in NWSRFS.
Brian Connelly will work with Janice Sylvestre to write a HOSIP proposal. This proposal will encompass both the threshold modeling approach and FLDWAV/SHRT enhancements to support river ice forecasting.
Other notes during the breakout session are as follows:
Dr. Steven Daly is planning to write up a followup proposal to refine the procedure to allow for forecasting low, medium, and high chances of ice jams, following the approach presented by Chandra Mahabir at the University of Alberta. This proposal will not address severity of ice jams, only probabilities.
It was noted that Rich Pomerlau at St.Paul Corps is computing AFDD for selected sites. The data is available on the St.Paul Corps website.
John Fulton of the USGS suggested the NWS talk to USGS district offices about getting ice discharge measurements for selected gages to assist our understanding of the affects of ice cover on river hydraulics. He suggested we get Emergency Managers and others to want the information also, as that will increase the chances of getting it. He also suggested we call the USGS offices to get the notes taken when the techs are in the field. Thomas Hawley reported that he has provided max/min thermometers for USGS sites that transmit with the stage data. He provided the equipment to the USGS and they installed it. Thomas will write up an informational document on how he accomplished this task.
The WFOs are concerned about how they will get the data and how it will be displayed on the AHPS website, but we decided we would figure that out once we were able to calculate the data at the RFCs.
Larry Rundquist is concerned about how we can forecast a stage during freezeup and with a full ice cover when we really don't know the ice-affected rating curve. The Canadians take stage data off their website when the gage is affected by ice and they put a comment "ICE" on the website instead. The WFOs would really like to have that capability. Michael Lukes said the IT at FGF has written a program that replaces the stage graphic on the AHPS page with a comment. The USGS has done studies on estimating ice affects but found it to be too difficult. Dr.Steven Daly said we could back calculate the ice affect if we know what the flow should be, and we know what the stage is reporting, we can back calculate to figure out how much ice is involved. WFOs think that people downstream would be interested in this type of information "Two and a half feet of ice affect."
Janice Sylvestere suggested we use HOSIP to get our requirements on the list and she also mentioned that the RFP deadline is December 15th for submitting Dr. Steven Daly's proposal. She also suggested we should get in on the agenda of a collaborators meeting somewhere and put a presentation together on data needs and resultant possible forecast information and see if we can get funding from other outside sources.