SUMMARY OF RESPONSES
THRESHOLD RUNOFF PLAN
1. The new ThreshR project is quite ambitious. [Eric Strem, CNRFC]
2. Is there a parallel project going on with a similar timeline to
address the modelling of soil moisture, ie a distributed Sacramento
Model or the like to provide soil moisture/runoff on a similar grid. [Eric Strem, CNRFC]
RESPONSE: We will continue to use the lumped SAC. There are no immediate
plans for a distributed model.
3. Can a WFO produce gridded precip without a radar? There are places
with marginal, at best, radar coverage and quite a few areas around
here where the radar overshoots the low based orographic
precipitation. This would make gridded precip based on the radar
questionable. [Eric Strem, CNRFC]
RESPONSE: Having gridded rainfall from the WSR-88D radars was the driving
force to provide gridded flash flood guidance. We can still provide
gridded FFG. But without radar coverage or marginal coverage, gridded
precip will not be available or only marginal. It's not evident how
gridded precips would be derived for these areas.
4. The assumption is that IHABBS and the underlying data are functional operational
programs. This is definitely not the case at NCRFC! We have attempted to put
IHABBS through its various incarnations and versions, and yet, we still cannot
produce useful accurate results. This is especially the case in the flat
topography of ND, where problems exist in removing non-contributing areas, and
in some cases tributaries are flowing into the wrong river system. I will not
get into a IHABBS bashing session here, simply put it doesn't work for us.
Our basin boundaries continue to be based largely on manually edited datasets.
Page 6, Section 5.2 refers to validation that occurred in the ABRFC area. It has
been an ongoing problem to extrapolate findings in one small area of the
country, and declare that a method or technique is sound. For example, WSR-88D
estimates continue to be a major headache in the northern latitudes. [John Halquist, NCRFC]
RESPONSE: It is too early in the development to provide a definitive solution. Our only
recourse for flat areas may be to use finer resolution data. ThreshR is being developed to use
finer resolution data.
5. OH's support of ArcView should not end with the initial purchase.
This is a national program, therefore continuing upgrades, tech
support, and training should be funded from the national level. [Bob Cox, MBRFC]
RESPONSE: Under the current plan OH will purchase (is purchasing) ArcView and
Spatial Analyst with upgrades. Tech support for the RFCs will come
from OH. For the second year OH will purchase upgrades for each RFC.
Beyond the second year the regions must provide the upgrades for their
RFCs. Tech support will continue from OH.
6. Training is included in the time line, but not mentioned. It
deserves more attention than that. [Bob Cox, MBRFC]
RESPONSE: I would agree that more details with regard to training
should worked out soon. Although use of the ArcView threshR
application will require only minimal GIS knowledge, I believe
that some basic training in GIS concepts and ArcView would be
highly beneficial. I have had several previous opportunities to
provide training on the subject of GIS and hydrology, and I hope
that this type of training may be provided in conjunction with
threshR training. Again, no formal plans have been worked out.
7. RFC coordination is a huge issue and always has been in regards to
the FFG program. How will we insure that RFCs are using consistent
methods? I have no objection to limiting RFC capabilities if it means
more consistency. [Bob Cox, MBRFC]
RESPONSE: A national spatial database will be provided with the
ArcView-based threshR. At a national level, this database will
support calculation of threshold runoff using the peakflow regression
approximation (i.e. Q2) for bankfull flow and Snyder unit hydrograph
peak flows (Method 4 from the Threshold Runoff Plan). Using this
baseline option, consistency could be achieved among RFCs. GIS
programming and training will be provided to support the use of other
methods, but it does not seem feasible in the short term to develop
nationwide data sets to support the other 3 methods described in the
Threshold Runoff Plan (also see Gregg's question 3).
8. The statement is made that, once threshR has been implemented, the
RFCs can begin issuing gridded guidance. This cannot truly occur
until the RFCs have the capability to implement distributed modelling
techniques. Where does OH stand on this? [Bob Cox, MBRFC]
RESPONSE: After threshold runoffs have been derived using threshR and
loaded into FFGS, gridded products are defined and then sent to your
WFOs via AWIPS. Zone guidance will be computed as you are doing
currently but using gridded runoffs from threshR. Gridded computations
remain as they are now, i.e., FFG at a grid is computed using the
gridded runoff and the rainfall-runoff conditions (by RFC forecast
basins) for that grid. FFGS will execute as it is now but gridded
products will also be generated.
9. Will we have the option of using the altered IHABBS flow-direction
grid? [Gregg Schalk, MBRFC]
RESPONSE: I am unclear on what you mean by the "altered IHABBS
flowdirection grid." Could you tell me what you mean by this?
What I mean with [the above] question is with the original Flow direction grid some
errors occurred mainly because of dangling streams. I believe the grids were
created to make sure they correspond with streams,. Other errors may also occurr
especially in flat terrain. We are able to correct that in IHABBS, thus making
a new flow-direction grid. Do we have the option of using the new grid instead
of the original grid created by NOHRSC? [Gregg Schalk, MBRFC]
RESPONSE: The short answer to your question is that you will have the capability
in Arcview to "correct" errors in the flow direction grid.
Unfortunately, you will not be able to use a previously corrected
NOHRSC grid because of map projection issues -- I have a long
explanation for this. I would very much like to discuss the specifics
of problems that you've had with dangling streams and flow directions
in the past so that we can minimize the need for duplicating any
previous edits that you have had to make to NOHRSC flow direction grid.
10. I would think we might want to be consistent in map overlays
(especially those given to the WFOs) of FFG basins, from arcview, and
NWSRFS basins, from IHABBS. Is there a possibility of different results?
RESPONSE: Yes, but not significant from a hydrologic point of view.
Are there different algorithms for basin delineation between IHABBS and ARCVIEW?
[Gregg Schalk, MBRFC]
RESPONSE: No, but the underlying data are slightly different due to the map
projection.
11. Why not use the unit hydrograph methods NOHRSC and company
developed or allow the option of using those methods? This would allow
us to have consistent regional parameter sets for both systems and not
require redundant efforts to develop regional paramenter sets. [Gregg Schalk, MBRFC]
RESPONSE: The unit hydrograph tool offered by NOHRSC includes several unit
hydrograph options. Each of these options requires some engineering
judgement in terms of manually entered parameters, which is probably
appropriate. These same methods can certainly be included as options in
the threshR application although the level of consistency in resulting
threshr data sets may depend more on engineering judgement than which unit
hydrograph method is used. As mentioned in the response to question 3
from Bob, it has been proposed as part of this project to do some research
and develop default maps of Snyder coefficients (nationwide if possible).
Regardless of the unit graph method used, having ArcView and the
appropriate geographic data sets will make visualization and storage of
unit hydrograph and threshr parameter sets easier.
12. Is there any effort in the future to validate threshR with real
data? [Gregg Schalk, MBRFC]
RESPONSE: There must be feedback from the WFOs to the RFCs. Over the
years we have become familiar with guidance based on the current
threshold runoffs. Generally, many of these values are
regional and don't reflect local hydrologic features. Runoffs from
threshR should vary considerably. Likewise the resulting gridded
guidance values will vary considerably from what we are used to
seeing. WFOs will undoubtedly question the new guidance but it can be
validated with feedback of flooding conditions from the WFOs.
13. Western Region cautiously approves of the latest plans for Thresh R. We are anxious to have
a tool which will be an improvement over our current flash flood procedures. At the same
time, we are concerned that the value of "gridded flash flood guidance" has been over-sold.
Here in the western U.S. (And probably elsewhere), we will be using crude approximations
of runoff potential and precipitation rates to produce such guidance. It will be important to
the proper usage of such any flash flood guidance that it be recognized that it is only an
approximation and not a substitute for paying attention to situations and using sound
judgement in analyzing and reacting to them. [Gregg Rishel, WRH]
RESPONSE: All our models are tools to help the professionals better analyze and act to
flood events.
14. More background on the justification for using the 15 arc second data is needed for
understanding why this resolution was chosen. As an example, we would like to see
explanations which would answer such questions as, "Will the 15 arc-second flood direction -
flow accumulation grids provide the detail needed to define basins in mountainous terrain?"
Such background should be added to the plan. [Gregg Rishel, WRH]
RESPONSE: There are certainly some questions as to how well the 15 arc-second data will
be able to define basin boundaries in both mountainous areas with narrow ravines (I assume
this is the problem you refer to) and flat areas. Although the use of higher resolution data is
not a specific part of the initial threshR work plan, the terrain analysis component of the
threshR programs will be designed to accommodate the use of finer resolution data if desired.
The initial effort is to implement a threshR tool that the RFCs will use immediately. Gridded
threshold runoffs derived from coarse resolution data is better than the runoffs currently in
use today.
RESPONSE: With regard to the use of higher resolution data, I have heard some claims that
the use of 3" data will provide no greater accuracy than 15" data in many cases; however, to
the extent that time permits I would like to examine some special cases of excessively steep
and flat terrain to see if any improvements can be achieved through the use of higher
resolution data. It would be helpful to me if you could identify specific locations where you
think the use of 15" data is inadequate.
15. The plan references a "database" that will contain the return flow equation parameters needed
to compute bank-full values. Geometric parameters are also to be created from IHABBS data
layers for each basin. Threshold runoff (local and accumulated) will then be calculated for
each basin. The plan is unclear on where these values generated from the geographical data
are to be stored. Will they remain in the ArcView database structure or is the desire to port
them to Informix? It seems that the intent is that this process will only need to be done once;
however, if basin definitions change (due to changes in gaging stations), additional data
becomes available, or current methods are modified, it would be beneficial to have this
information stored in one location for rapid updating and retrieval. If this is the intent of the
plan, it would be worthwhile to examine the abilities of the Spatial Database Engine (SDE)
for Informix. [Gregg Rishel, WRH]
RESPONSE: The gridded threshold runoffs from threshR will be loaded into the Flash
Flood Guidance System (FFGS). The FFGS database will be moved into Informix when the
NWSRFS database is moved into Informix as part of IHFS. The threshold runoffs will only
need to be derived once. As finer resolution data becomes available or new methods are
implemented, threshR will need to be re-run.
RESPONSE: I agree that it might be worthwhile to investigate SDE. Although, I don't
know any details about SDE, I am familiar with the concept. In the meantime, I don't think it
will be very labor intensive or computer intensive to transfer new data from ArcView to FFG
if basins are redefined.
16. We view the abandonment of GRASS and move to ArcView as a very positive step. Using
ArcView utilities will make this process more user friendly. An explanation of how
IHABBS and the gridded fields developed for IHABBS fit into the Thresh R basin
delineation procedures is needed. The present plan is unclear how using the IHABBS data
and Arc View will fit together. [Gregg Rishel, WRH]
RESPONSE: Since ArcView must work with projected data, an appropriate database must
be derived from the IHABBS database. ArcView will not access IHABBS data but a data set
derived from IHABBS.
17. One area we would like to see expanded is the capability to select a flow regime other than
Mannings flow equations or a 2-yr return period for calculating threshold runoff values.
Data for Mannings flow equations will not be available for many areas of the west.
Furthermore, a 2-yr return period flow is simply not a good indicator of bankfull conditions
in many areas of the western U.S. We feel that the ability to select return period of other than
2-yr is an essential and significant component of making Thresh R flash flood guidance as
representative as possible in the western U.S. [Gregg Rishel, WRH]
RESPONSE: Limitations of the 2-yr return period flow has been addressed in Section 9.0 of
the Plan. Regression equations will be available for 2-, 5-, and 10-yr return periods.
18. Western Region suggests that the implementation of Thresh R be linked to a verification plan
which would provide feedback on the validity of the FFG values it produces. Such feedback
is needed in order for Thresh R to "earn its way" into use at the WFOs. Verification would
also be useful in quantifying the usefulness of NWS Flash Flood products. [Gregg Rishel, WRH]
RESPONSE: The WFOs must provide the feedback to verify FFG values. The feedback
will reflect the validity of the threshold runoff values as well as the performance of the soil
moisture model used by the RFCs. If the RFC models are kept balanced to track the rainfall-
runoff events for regular hydro forecasts, then any discrepancies between FFG values and
flash flood events can be attributed to threshold runoff values.
19. The plan does not include any discussion of proposed training for the RFCs in implementing
Thresh R. Similarly, there is no mention of training for the WFOs in understanding Thresh R
flash flood guidance limitations and proper utilization. Training in both of these areas will be
essential to the successful implementation and usage of Thresh R flash flood guidance. A
suggested approach for RFC training would be for OH to hold a Thresh R workshop in each
region. Training for the WFOs in the usage and limitations of Thresh R could be distance-
learning based such as a web site or CBL with reinforcement from the servicing RFC.
[Gregg Rishel, WRH]
RESPONSE: There must be training for use of threshR at the RFCs and use of gridded flash
flood guidance at the WFOs. Gridded FFG is used in WHFS which will need to include
training on use of gridded FFG. WFOs will not normally see the threshold runoff values but
they will see the gridded FFG values computed from the runoffs and current soil moisture
states. As the technical portion of threshR is developed, we will get a better indication of
what training is needed.
|
Follow us on Twitter
Follow us on Facebook
Follow us on YouTube
OWP RSS Feed