Klamath Basin Water Crisis
Upholding Americans' rights to grow food,
own property, and caretake our wildlife and natural resources.
September 5, 2003
Klamath Basin Rangeland Trust
Response to comments provided by Mr. Bartell
KBRT responses are embedded within the original document. The responses are indicated by green italicized font.
Comments on KBRT Monitoring Plan, and Last year’s Work
Questions for KBRT?
The color aerial photos in the report were flown by WAC, Inc for Graham Matthews & Associates (GMA) on August 8, 2002. GMA has had a number of aerial photo sets flown for various parts of the Wood River Basin over the last 6 years mostly concentrating on the lower Wood River, Crooked Creek, Fourmile, and Sevenmile Creek channels. Nearly all of the Pilot Project area is covered by air photos, however, complete low-altitude coverage could only be obtained from air photos taken over the course of several different years. There is complete coverage at 1:24000 of the Wood River Valley from WAC in the last few years.
These following are initial comments on the KBRT monitoring plan; more comments may be made after it is reviewed further, or upon replies from KBRT.
One main overlying comment is that there is a crucial lack of explanation of what is going to be done with the data after it is gathered, and how conclusions will be drawn. All of the proposed methods for determining if water is gained and nutrient loading should be identified in detail. In a similar format to this; "the stream flow has been measured at Dike Road for x number of years by BOR. This data will be compared with data at Weed Road to determine if KBRT increased flows, and the process for analyzing and comparing the data is explained here."
After the surface water data has been collected it will be analyzed in a synoptic manner. Discharge records will be compared to pre-existing data on a site to site basis, where possible, in order to determine if water is gained through the actions of KBRT. This is not always easy to do as variations in streamflows are naturally occurring from year to year. As more data are gathered it will become easier to notice and define trends.
2.0 Surface water monitoring, much of the monitoring appears to be baseline data (whereas there is not a period of record to compare with). This data will be worthless if a period of record cannot be established in the future to monitor year to year trends and changes. More permanent gauges should be established, with all information publicly available and managed by a third party government agency such as USGS, with adequate funding for them to do so on a multiple year basis.
All of KBRT’s gages will be operated on a long term basis so long as funding is available. All data collected by KBRT are public and available to anyone by request. Summaries of all of the data are provided in our reports, and any raw data can be supplied if a specific and detailed request is made. The volume of raw data collected precludes inclusion of all of it the annual reports. All of the work done by consultants on behalf of KBRT is reviewed by the USGS as well as professionals from other agencies. If the USGS is given sufficient funding to complete monitoring on a similar scale to that currently conducted by KBRT, and is able to meet the same timelines for data analysis and presentation that KBRT is held to, we would be supportive of the USGS taking over this work on our behalf.
Monitoring surface return flows, (as identified on the monitoring report page 3), is not an accurate accounting of total return flows to the system, whereas, much of the water will return to the system subsurface, either increasing flows of existing springs or otherwise flowing in the riparian area or in the bed of the stream.
We are not completely clear on your comment, but if you are referring to the fact that if you irrigate a field some of the water will be returned to the stream channel by other means than return flow ditches, we agree. Some irrigation water will be returned to the system subsurface or will be returned to the system in a non concentrated fashion that makes it difficult to measure as a point source return. In order to account for this, we established streamflow accounting units on Crooked Creek in order to determine if return water was entering the system that was not directly measured in a concentrated point source. The accounting units on Crooked Creek are explained in detail in Section 4.1.5 of the 2002 pilot project monitoring report. In addition, a central purpose of the groundwater monitoring work is to assess the magnitude of groundwater return flows.
On page IX of the 2002 Monitoring Report it is stated "Modeling of the shallow groundwater system also indicates that a significant portion of the irrigation return flows do not reach Agency Lake during the irrigation season. The model results suggest that in addition to the water lost from the system as the result of evapotranspiration (see below), at least 10% of the irrigation water diverted does not return during the irrigation season due to delays in return flows. If a historic diversion rate of 6.5 feet is assumed for the 2002 pilot project area, the model results suggest that 2050 ac-ft of irrigation water historically did not return to Agency Lake during the irrigation season as a result of the delays in return flows"
The above statement assumes two things, A) That irrigators diverted 6.5 acre feet per acre prior to last year’s KBRT water marketing on 3150 acres. B) That 10% of the water diverted never returns to the stream. Neither of these statements can be supported by scientific fact and should be removed from the report.
As described above this assumption was never made, so the relevance of this question is unclear.
The model is nothing more than a series of hypothetical calculations which will be incorrect by a order of magnitude if incorrect or unknown data is inputted, as was done here, and noted above. Modeling is highly subjective at best. When models are not ground truthed to a substantial data set, they are nothing more than guessing.
Groundwater models are commonly used in two ways:
Neither of these functions constitutes "guessing" when the work is conducted by professionals using judgment, even if field data is limited to the most basic mapping, hydrologic, and geologic inputs. Variables used in the 2002 analysis included geography (well known), major hydrologic features (well known), elevations of secondary ditches (not known absolutely, but roughly known in relation to land surface), soil permeability (measured and constrained by mounding analysis, but likely variable), effective aquifer thickness (not well known and likely variable), and percolation quantity (not well known, and used in sensitivity assessment).
The 2002 KBRT groundwater model was used in both ways listed above. It was used as a guide to qualitatively understand the groundwater return flow component (#1 above), to plan 2003 field work (#1 above), and to estimate the return flow quantity and timing (#2 above). A high degree of confidence was placed on the qualitative finding that subsurface return flows are substantially delayed, and on our improved understanding of the variables that influence the quantities and timing. Because of the dependence of the quantitative results on (reasonable but) estimated variables, less confidence was placed on the quantitative predictions. Different words were used to describe the various model findings, depending on the degree of confidence for each finding.
The analysis shows that timing of subsurface return flow is an important variable in evaluating the effects of irrigation, and the 2002 KBRT model remains the most advanced tool available to address the issue. Over time, the more numerous monitoring locations and longer monitoring records can be used to improve modeling, and may in some cases supplant this and other models as the best basis to understand the basin hydrology. However, it is likely that modeling will remain a powerful tool for understanding the complex relationships between surface water/groundwater.
Ground water monitoring is in essence saying this tiny hole represents thousands of acres, which in many cases it will not. There are many areas of water upwelling, and sub-irrigation out of rivers stream and canals, in the Wood River Valley, where water will be made available to plant life through out the season, including areas on this property.
Water levels in piezometers and wells reflect groundwater head, which typically varies smoothly from location to location according to influences of recharge, discharge, and permeability. Such smooth spatial trends are confirmed by mapping out heads from a number of monitoring wells completed in the same water-bearing unit. Thus, although the measurements are at a certain location, head between the monitored locations can also be inferred.
In general, the deeper the aquifer, the more smoothly the head will tend to vary both in space and time. For instance, the USGS appears to monitor only five wells (quarterly) in the Wood River Valley for target aquifers at 92 to 226 feet depth. The shallow groundwater that is of primary interest to KBRT will be more variable because of the influences of drainage features and irrigation. In order to properly address variability in the shallower zones, KBRT monitored 13 piezometers in 2002 and is currently monitoring 17 piezometers plus one deeper well. Hand-measurements are collected weekly during most of the year, and several locations have loggers that record water levels several times a day. While the anomalies cited by the commenter may very well exist, the groundwater monitoring data probably document most of the range in groundwater depths and responses that occur in the valley.
Additionally there is an admitted sampling bias of not sampling peat soils because of a fear of fire, and other likely sampling biases of convenience to access. A process of arbitrarily selecting sites to drill holes, is not scientifically valid since sites will vary greatly depending on their proximity to wetlands, drains, ditches remaining full of water for stock, irrigation canals, river systems and etc.
The sites selected for groundwater monitoring were not random. As was described to the commenter during the science review team meeting, the sites were selected to create 3 transects across the valley to the best degree possible given property ownership/access issues as well as proximity to irrigation canals and rivers. The transects for the 2003 monitoring season include a north-south transect along the east side of the valley, a north-south transect along the west side of the valley, and an east-west transect across the center of the valley. Some of the sites selected are purposefully located adjacent to irrigation ditches in order to evaluate the radius of influence from these ditches on the groundwater table. In addition piezos 8 and 9 are located in peat soils, with P-8 located adjacent to an undrained wetland area that is adjacent to a non-irrigated pasture. In 2003 piezo 17 was added in a drained wetland area that is enrolled in KBRT.
There is a clear lack of data to support any numbers associated with a groundwater gain to the stream. In order to develop data, which would support a quantifiable rate of water gained via groundwater, hundreds of shallow wells would have to be drilled in a uniform and surveyed grid pattern throughout the area. Additional tests would have to be done as well, such a placing dye in the water to track the source and movement to the river.
Evidence of groundwater discharge to the streams and rivers of Wood River Valley does not require the effort or type of analysis suggested by the commenter. Groundwater discharge is commonly identified by changes in streamflow, stream temperature, stream chemistry, and by documenting groundwater elevations above surface water elevations. Indirect methods are commonly used because groundwater discharge is usually diffuse and while important from a groundwater perspective, it may constitute a small percentage of surface water flow, and thus not be directly measurable via stream gaging.
Even without direct examination, groundwater discharge to the surface waters is evident considering the spring-fed origins of many of the creeks, and the fact that some wells have heads above ground surface.
This lack of data is clearly outlined in the KBRT report with such statements as Page C-2 "…the actual rates of recharge from irrigation are poorly documented,.." Page C-5 "One simplifying assumption of the model is that other sources of recharge are not simulated" C-7 "Consideration of natural recharge could also affect the partition between subsurface infiltration and rejected recharge,.." Page C-7"…it (the model) has not been formerly calibrated to the project site and may not represent al the processes occurring onsite. Specifically, the magnitude of irrigation application and the process by which the divide between infiltration, subsurface flow to shallow drains and surface-water runoff is not well understood"
I fully understand KBRT has a large but limited budget and I understanding all these processes would cost millions of dollars, however the conclusions should never extend beyond the quality of the data. Unless the millions of dollars are spent to understand all these processes no attempt should be made to quantify a specific number attached to water returning via groundwater, utilizing groundwater monitoring, or ground water modeling. Instead KBRT should only assume limited generalizations from the very limited data they are collecting compared to fully understanding the whole picture.
The statements that the commenter quotes above to indicate the limitations of the KBRT dataset are all valid and were included in our report so that the reader would understand the limitations of our data set. KBRT fully concurs that additional data beyond the partial year of information collected in 2002 are necessary to evaluate the potential benefits from the KBRT management plan. The meaning of your comments is unclear since any limitations of our dataset are described multiple times in our 2002 report. In hydrologic work, modeling and data collection are always considered iterative processes, where initial modeling work helps to define the important variables, additional monitoring provides information about those variables, which feeds back into the model, and the process is repeated. This method is generally applied both to actual models, as well as conceptual models about system function.
With respect to the groundwater modeling, the conclusion of the current model (delayed baseflow) has high confidence in a qualitative sense, and is the best quantitative estimate currently available. The quality of this estimate will improve with current and future data collection and model refinement.
Additionally a total amount of water gained can be determined with proper flow monitoring (of surface water) at much less cost.
This comment is unclear as it conflicts with the commenter’s previous comment that, "Monitoring surface return flows, (as identified on the monitoring report page 3), is not an accurate accounting of total return flows to the system, whereas, much of the water will return to the system subsurface, either increasing flows of existing springs or otherwise flowing in the riparian area or in the bed of the stream." Inclusions of groundwater monitoring, as well as evapotranspiration monitoring, is necessary to fully understand the water budgets for this hydrologic system. Furthermore, streamflow accounting alone can only indicate the total stream gain in a given reach, but cannot determine if the gained water is from natural discharges to the stream, or if it is subsurface irrigation return water. As a result, it is important to utilize groundwater monitoring and modeling techniques to assess and seperate the groundwater inputs from natural and irrigation sources.
The implication is made that dropping the water table will make the water unusable by plants for evapotranspiration. While the two are connected, the soil can wick water a considerable distance to the surface, or soil moisture can remain as the water table drops. Soil moisture and the color of the foliage should be utilized to determine whether the plants are taking water instead of the water table.
Soil moisture in usable quantities is found above the groundwater table as is discussed in our analysis. However the distance above the groundwater table that water can be held is limited, and depends on a variety of factors including soil type, and evapotranspirative loss from the upper portion of the soil column as the growing season progresses. Monitoring of the groundwater provides important information about potential water availability to the plants given these parameters. In addition, KBRT has added soil moisture probes to our monitoring during the 2003 season since we have sufficient funds to include it this year. In 2002, we did not have sufficient funding to monitor both, and so the groundwater piezometers were used since information for multiple aspects of the monitoring was available from them.
Small irrigation test plots, as identified in the monitoring report page 14, should not be used since much more water will be needed to supply the surrounding non irrigated lands, and the irrigation rotation will not match other areas in the Wood River Valley.
The irrigation test plot is a small parcel located within a much larger irrigated pasture, not a non-irrigated pasture as the commenter suggests. The grazing and irrigation patterns on this property are typical of those throughout the Wood River Valley.
The sites selected for the ET stations were chosen to be as representative as possible of conditions in the valley. The sites are located near the Wood River so that if non-irrigated areas near a river have greater water availability to the plants as you have suggested, this will be accounted for. The sites were also placed in the center of the valley so that climatic parameters will be representative of the largest areas of the valley possible. In addition, a third station has been installed at the north end of the valley since the northern area typically receives more precipitation and more cloud cover than the south end, and therefore may have lower ET rates than the center or south end of the valley. In addition, at least 10 Hobo weather stations are operated in transects across the valley to evaluate any variations that may occur near the valley margins, or near the lake.
The purpose is to gather baseline data with respect to KBRT activities. As we collect data over time, we will compare back to our own data.
Serious problems with KBRT monitoring.
As described above, conditions in the drained wetland areas, and in areas adjacent to creeks and irrigation ditches are being evaluated. No lands that have not been effectivelyirrigated in the past, and no presently existing wetlands, are included in KBRT leases.
Wet soil (class IV) was recently reassessed by the Klamath County assessor in the Wood River Valley at a devalued rate because "it was boggy", and the water could not be removed from it causing a lesser quality of forage.
All lands enrolled in KBRT are dry and can have water removed from them.
KBRT assumes an across the board ET rate for all lands under contract, historic maps clearly mapped wetlands, these lands should be excluded from areas where water is assumed to be gained from. To the extent pumping stations are stopped or drainage ditches filled there should be a depletion in flow assumed from these lands.
Again, no wetlands are included in KBRT. Only lands which can be kept dry throughout the season are included.
Many people have observed that much of the KBRT land remained green throughout most of the irrigation season last year, this is very indicative that much of the vegetation is obtaining water subsurface resulting in minimal increases in flow.
The ET monitoring that is currently being conducted will assess the volumes of water lost to the atmosphere under irrigated and non-irrigated conditions. However, some basic logic can also be used, if the grasses are able to obtain sufficient water without irrigation, then why do ranchers in the valley spend so much time, energy, and money on irrigation? In addition if the grasses get sufficient water without irrigating, there would be no reason to file for water rights or spend time going through the adjudication process.
2) Monitoring last year appears to be full of apples and oranges comparisons i.e. flows from the 1960’s on an upper reach were compared to flows from not only different years but different centuries at a lower reach (the 2000’s)(figures 3-5, 3-8 and page viii). The difference in flows over this period are likely due to less water entering crooked creek from the now abandoned portion of the Fort Creek ditch and different irrigation patterns not anything KBRT has done.
Figure 3-5 is not a figure describing flows. This figure is a plot of gage height files collected at the Root Ranch from 1998-2002. The Root Ranch is well within the backwater effect of the lake so comparisons of gage heights are not a reliable indicator of discharge. Figure 3-8 is a figure describing flows and can be considered an appropriate tool for comparing flows on Crooked Creek. While the data collected by the USGS was not at the same exact location as the current gaging station on Crooked Creek it was very close. The purpose of the analysis here is to begin to understand how the cessation of irrigation has affected flows in the Crooked Creek System. It is specifically stated in the text of the report (Appendix A, pg A-11) that estimating increased flows due to the KBRT project is very complex due to the fact that individual water years are not strictly comparable due to annual changes in discharge rates from the springs. One of the assumptions stated on Page 9 of the 2002 Pilot Project Monitoring Report states that it was assumed that irrigation practices have remained relatively consistent along Crooked Creek since the mid 1960’s. If there is evidence that that irrigation practices have been substantially changed by landowners since that time we would like to have that information relayed to us so that we may re-evaluate our analyses.
Monitoring fails to establish changes in nutrients flow or temperature on a year to year basis or pre and post KBRT project i.e. monitoring of fish and invertebrate was done in Crooked Creek, they found healthy populations. However, they may have always been healthy and in fact likely were, since they probably would not have appeared in such a short period since the last irrigation season. This really tells nothing about the effect of the project.
We were not attempting to establish changes in nutrients or temperature on a year-to-year basis. The data contained in the report is only the beginning process of establishing baseline data and only include data for a partial season. The goal is to monitor the response of the system to KBRT activities. This will be accomplished by monitoring the system over time. If a system is already healthy, then presumably we would not see a response to KBRT activities. The fish data for Crooked Creek do not indicate anything about the health of the population.
It appears the piezometer sites were arbitrarily selected and compared to sites in a completely different area.
The site selection methods for the piezos are described above.
3) KBRT is seeking to prove it is doing a variety of good things, rather than trying to determine what is happening in the ecosystem. By putting forth a series of hypothesis, KBRT admits it has a bias toward a pre-determined conclusion. This presents a major problem because earth science is very subjective and conclusions can be greatly varied by what data is inputted.
KBRT is not conducting the monitoring work, but has instead hired consulting companies with professional scientists to conduct the work. All of this work has been reviewed by the USGS and scientists from other federal agencies. As stated above, if the USGS can conduct a similar level of monitoring on similar timelines to the work KBRT has contracted from private consulting companies, KBRT would be supportive of USGS operating the monitoring program.
In order to assure there is no bias in the work, the monitoring moneys and monitoring activities should be turned over to a nonbiased third party (such as USGS), with no ties to the principals of KBRT or any other financial stake in the outcome of scientific monitoring.
See comment above regarding scientists and USGS involvement. KBRT is a public purpose non-profit organization and cannot realize any financial return from the project.
4) The Data from last year does not support the conclusions; in a letter dated February 13, 2003 on behalf KBRT Kurt Thomas states that "Crooked Creek stream flows were increased by 10-15 cfs on average relative to available historic discharge data throughout the period May 1 to October 31". Yet when one looks at the flow data (figure 3-5) one finds that Crooked Creek had one of the worst flow on record in fact flows were worse than in the severe drought of 2001. The only year that had worse flow was in 2000. In talking to locals it appears the 2000 low flow was due to a supply canal to crooked creek being washed out, this is supported by the anomaly on the hydrograph. Given KBRT’s own data it appears one could reach a completely opposite conclusion being that KBRT is depleting flows.
KBRT should provide the precise scientific calculations utilized to support the 10-15 cfs gain, in a manner that can be replicated by other scientists. If the data cannot be replicated by scientifically viable methods the 10-15 cfs gain should be removed from the report.
Figure 3-5 does not present flow data, but rather presents gage height records from Crooked Creek at Root Ranch. Gage height is not a reliable indicator of discharges at the Root Ranch gaging locations due to the fact that this site is well within the backwater effect of the lake. Gage height at the Root Ranch is influenced more by lake level than it is by streamflow. In 2002, a Starflow Ultrasonic Doppler Current Meter was installed at the site in order to accurately measure discharge. No continuous discharge records have been computed for the years of 1998-2001.
On page viii of the Executive Summary in the 2002 Pilot Project Monitoring Report it is specifically stated that the 10-15 cfs increase in discharge in the Crooked Creek System is a preliminary estimate and that additional data are necessary to confirm this estimate. The 2002 Pilot Project Monitoring report (Page 9, last section) describes in detail the 3 methods used to derive the average 10-15 cfs increase in discharge on Crooked Creek due to the cessation of irrigation in the pilot project area. While each method of analysis produced a different value for the estimated increase in streamflow on Crooked Creek 10-15 cfs seems to be a reasonable value.
Page Updated: Thursday May 07, 2009 09:14 AM Pacific
Copyright © klamathbasincrisis.org, 2003, All Rights Reserved