Discussion of Results

The results from this study showed that concentrations of nutrients in the entire lower Yakima River were high enough at certain times and places to support the abundant growth of phytoplankton, periphyton, and macrophytes and that the metabolism associated with this plant growth caused exceedences of the Washington State water-quality standards for dissolved oxygen and pH. The abundance and distribution of these aquatic plants, however, varied greatly throughout the lower river and, in the case of macrophytes, between years.

Comparison of Nutrient and Algal Biomass Conditions to Suggested Criteria

The comparison of nutrient and algal biomass conditions to suggested criteria were based on samples collected between July 1 and September 30, 2004–07. All concentrations of dissolved nitrite plus nitrate and total nitrogen measured in the Mabton and Kiona reaches were greater than the suggested USEPA reference conditions. In contrast, the concentrations of dissolved nitrite plus nitrate and total nitrogen measured in the Zillah reach often were less than the reference conditions and the mean concentrations of these nutrients measured in the Zillah reach were much closer to the reference conditions compared to the Mabton and Kiona reaches (table 12). Although all concentrations of total phosphorus measured in the Kiona, Mabton, and Zillah reaches were greater than the reference conditions, the mean concentration measured in the Zillah reach was much closer to the reference condition compared to the Mabton and Kiona reaches. One-quarter or less of concentrations of dissolved nitrite plus nitrate, TKN, and total nitrogen, and one-half of concentrations of total phosphorus measured in the Naches River, were greater than the reference conditions and the mean concentrations of all four nutrients were less than the reference conditions.

Twelve percent of the phytoplankton Chl a concentrations measured in the Yakima River and no phytoplankton Chl a concentrations measured in the Naches River were greater than the State of Oregon nuisance level of 15 µg/L. Thirty-nine percent of the periphyton biomass samples in the Zillah reach, but no periphyton biomass samples in the Naches River or the single periphyton biomass sample in the Kiona reach were greater than the Province of British Columbia nuisance level of 100 mg/m² (table 12).

The mean concentrations of nitrite plus nitrate, total nitrogen, and total phosphorus measured by WA DOE in the Zillah reach (Washington State Department of Ecology, 2008b) were less than the concentrations of nutrients measured in the reach for this study and, unlike the concentrations measured in this study, were close to or less than the reference conditions (table 12). This was most likely because the WA DOE samples were collected at a fixed station (RM 111) upstream of major nutrient inputs to the river—the City of Yakima wastewater treatment plant, Wide Hollow Creek, Moxee Drain, and Ahtanum Creek, which together contributed about 33 percent the total nitrogen load and 47 percent of the total phosphorus load that entered the Zillah reach during the 2004 synoptic survey, whereas almost all samples for this study were collected downstream of these inputs.

Factors Related to Periphyton Biomass

The lack of a positive relation between surface-water nutrient concentrations and algal biomass in the Zillah reach could mean that nutrients were being removed from the water by algae or that physical factors and (or) grazing by invertebrates, rather than nutrient concentrations, were controlling algal growth. The shallow and relatively wide channel in the Zillah reach may have provided a large amount of habitat favorable for algal growth where light was available for photosynthesis. Moderate-to-high water velocities in this reach may have favored enhanced growth by delivering nutrients or removing waste products at higher rates (Horner and others, 1983; Stevenson, 1997). Higher water velocities, however, result in increased drag on algal cells, which can reduce biomass (Stevenson, 1997). This may explain the higher biomass observed along the stream margins in the Zillah reach, where the water velocity was less than in the main current.

The lack of a positive relation between surface-water nutrient concentrations and algal biomass in the Zillah reach also could have been due to the complicating effect of ground water. The paired pore water and surface water nutrient and hydraulic measurements in the Zillah reach in 2006 and 2007 showed that the infusion of nutrient-rich ground water may have stimulated periphyton growth in some sections of this reach. These results are consistent with the findings of other regional studies. According to Kinnison and Sceva (1963), “A large amount of subsurface return flow from irrigation enters the river between this station (Yakima River near Mabton at RM 55) and the next upstream at Parker (RM 104).” Previous studies measured nitrate concentrations between 1 and 5 mg/L in about 250 of nearly 500 wells sampled in the watershed for Toppenish Creek (which enters the Zillah reach at RM 80), with nitrate concentrations from a few wells reaching 5–10 mg/L (Sumioka, 1998).

Results from the 2004 synoptic survey indicated that regional ground water likely was supplying nutrients to the Zillah reach. The nitrogen load from the small irrigation return drains between RM 86 and 83 could not account for the sudden increase in the concentration of dissolved nitrite plus nitrate between these two locations. The sudden increase could have been caused, however, by upwelling of nutrient-rich ground water related to the regional geology. The lower Yakima River basin consists of a series of troughs and ridges that run perpendicular to the river. The river generally loses water at the upper ends of these troughs and gains ground water at the lower ends. The Yakima River gains water between RM 85 and RM 81, where Toppenish Ridge and Snipes Mountain (which are part of the same formation) intersect the river and form the lower end of one of these troughs. Some of this gain can be attributed to the release of ground water from the alluvial aquifer as it pushes up against the bedrock outcrop running along the northeast side of the river in this area (John Vaccaro, U.S. Geological Survey, written commun., 2008).

Factors Related to Macrophyte Biomass

The most prolific macrophyte in the lower Yakima River was, by far, water stargrass, especially downstream of Prosser Dam at RM 47. This plant is found throughout the United States (Horn, 1983; Smart and others, 2005) and is native to Washington State. Although its great abundance in the lower Yakima River poses a nuisance at times by interfering with recreation and contributing to low dissolved oxygen concentrations and high pH, moderate amounts of water stargrass provide fish habitat and waterfowl food, and the plant has been used for river habitat restoration by the U.S. Army Corps of Engineers (Smart and others, 2005). Abundant growth of water stargrass in the Yakima River, however, can provide habitat for introduced fish species such as largemouth bass that can prey on small salmonid fishes.

Water stargrass typically grows in shallow water up to 1 m deep (Hamel and others, 2001), although it tolerates decreased water transparency (Davis and Brinson, 1980) and has been observed at depths of more than 4 m in lakes (Davis and Brinson, 1980). The roots are fibrous and grow from rhizomes in and on the riverbed (Hamel and others, 2001). The root systems of water stargrass beds in the Yakima River extended well into the river bed. Water stargrass may overwinter as a dormant root crown (Smart and others, 2005) or, if conditions permit, as entire plants (Horn, 1983). Growth begins in spring when water temperatures exceed 8°C and ends when water temperatures fall below 10°C (Horn, 1983).

Differences in substrate stability may account for the substantial differences in macrophyte biomass between the Zillah, Mabton, and Kiona reaches in 2005. In the Zillah reach, the river flows on alluvial and glacial deposits, allowing the river channel and its cobble substrate to move more readily than in the Kiona reach, where much of the river is constrained by bedrock. In the Mabton reach, the soft unconsolidated mud substrate moves even more readily during high water events than the cobble substrate in the Zillah reach (Jim O’Connor, U.S. Geological Survey, oral commun., 2007). As a result, macrophytes in the Zillah and Mabton reaches may not have a stable base on which to establish and expand year after year.

The large differences in macrophyte biomass between the Zillah, Mabton, and Kiona reaches in 2005 could not be explained by reach-scale differences in light availability, velocity, or water temperature. Although light was insufficient for macrophyte colonization more often and in more locations in the Mabton reach compared to the other two reaches, light penetration generally was adequate for plant colonization in all three reaches of the river. Velocities were similar in the Zillah and Kiona reaches but macrophyte biomass was much higher in the Kiona reach, and the water temperature in all three reaches was warm enough throughout the growing season to support macrophyte growth.

The large differences in macrophyte biomass between reaches in 2005 also did not correspond to reach-scale differences in water-column or bed-sediment nutrient concentrations. Although the median macrophyte biomass in the Zillah and Mabton reaches were equal to 0 g/m², the concentrations of bioavailable nutrients in the Mabton reach were 4 times greater than the concentrations in the Zillah reach, and although the concentrations of bioavailable nutrients in the Kiona reach were comparable to the concentrations in the Mabton reach, the macrophyte biomass in the Kiona reach was much higher. A previous study determined that the median concentration of total phosphorus in the Quaternary deposits underlying the Zillah reach was similar to the median concentration in the Columbia River Basalt group rocks underlying the Kiona reach (0.10 and 0.09 percent, respectively) (Fuhrer and others, 1994); another study determined that the mean concentration of total phosphorus in the sediments of the Yakima River was 0.096 percent (Carlile and McNeal, 1974). The latter study, however, did not compare the longitudinal differences of biologically available forms of phosphorus in Yakima River sediments.

The gradual increase in macrophyte biomass in the Kiona reach between 2001 and 2005, followed by an abrupt decrease in 2006 and 2007, was best explained by changes in turbidity and streamflow (both of which influenced light availability) during that period. The number of days with above-average spring streamflow in the Yakima River declined from 1997 through 2005 including droughts in 2001 and 2005. Spring streamflow in the lower Yakima River is dominated by snowmelt from the Cascades, and below-average snowpack was measured during 4 of the 6 years between 2000 and 2005. Modest or low spring runoff would have decreased turbidity, depth, and velocity and led to improved growing conditions for macrophytes in the spring.

Implementation of improved irrigation practices by growers in the lower Yakima Valley during the 1990s decreased erosion and reduced the amount of sediment entering the Yakima River, causing a reduction in the loads and concentrations of suspended sediment during the irrigation season at Kiona between 1995 and 2003 (Coffin and others, 2006). The median turbidity at Kiona during the irrigation season decreased between 1995 and 2005 (Washington Department of Ecology, 2008). The decrease allowed sunlight to penetrate deeper through the water and create conditions more favorable for macrophyte growth throughout the irrigation season.

Increased spring runoff in 2006 and 2007 due to higher snowpack in the Cascades led to conditions that were not as favorable for macrophyte growth. The river generally was more turbid, deeper, and faster from March through June during 2006 and 2007 compared to 2005. Between March 1 and June 30 in 2005, 2006, and 2007 the median turbidity at Kiona was 4.6 FNU, 18 FNU, and 15 FNU, respectively. The median turbidity exceeded 10 FNU on 7.3, 66, and 54 percent of the days between March 1 and June 30 in 2005, 2006, and 2007, respectively. The mean gage height at Kiona during the same period was 3.8 ft in 2005, whereas it was 6.3 ft in 2006 and 6.6 ft in 2007. During routine streamflow measurements at Kiona the velocity in the river was lower during spring 2005 than in 2006 and 2007 (Gregory Ruppert, U.S. Geological Survey, written commun., 2008), which may have allowed plants to grow in more locations across the cross section in 2005.

Physical Factors Related to Dissolved
Oxygen and pH Conditions

Any complex dissolved oxygen model developed for the lower Yakima River would need to include the parameterization of the relation between dissolved oxygen concentrations and the respiration associated with autotrophic and heterotrophic metabolism. Because the parameterization of this relation might be difficult, understanding the relation between daily minimum dissolved oxygen concentration and a parameter that is relatively easy to predict and might be related to respiration, such as water temperature (Hill and others, 2000), could be useful.

The daily minimum dissolved oxygen concentrations and water temperatures at Kiona, Mabton, and Zillah occurred in the early morning. This pattern was opposite of the expected pattern from temperature-controlled equilibrium (higher dissolved oxygen solubility with lower water temperature), and was due to respiration by aquatic plants and animals. The daily minimum dissolved oxygen concentration at Kiona, Mabton, and Zillah also was strongly and negatively related to the maximum water temperature on the preceding day (table 13). The explanatory power of this linear relation did not increase when other factors such as streamflow and turbidity were included. A weak linear relation (r² ≤ 0.23) also was noted between daily maximum pH at Kiona and maximum water temperature.

Although the dissolved oxygen - temperature model was able to reproduce the general patterns in daily minimum dissolved oxygen levels, it tended to have a positive bias in the year with the greatest macrophyte growth (2005) and a negative bias in the years with relatively low macrophyte growth (2006 and 2007) (fig. 27). For some periods in each year, however, the fitted values were within the allowable error for dissolved oxygen readings (0.30 mg/L). The RMSE value in figure 27 represents the mean distance of the measured data from the fitted line (expressed in milligrams per liter) and evaluates how well the model fit the measured data for each year. On the basis of RMSE values the model was most accurate in 2007 and least accurate in 2005. The strong relation between the daily minimum dissolved oxygen concentration and water temperature could prove useful if a dissolved oxygen predictive model is developed for the lower Yakima River, and it also could be used as a simple stand-alone model for assessing water-quality conditions.

The summer monitor deployments in the Zillah reach in 2006 and 2007 with the highest water temperatures tended to have a higher percentage of time when the concentrations of dissolved oxygen were below the Washington State standard, whereas the opposite was true for the monitor deployments with the lowest water temperatures. These results were consistent with the strong relation between water temperature and minimum dissolved oxygen concentrations during the seasonal monitor deployments at Kiona, Mabton, and Zillah. There appeared to be no relation, however, between water temperature and pH during the 2006 and 2007 monitor deployments in the Zillah reach or the Naches River.

Factors Related to Spring Gross Primary Productivity at Kiona

The relations between the spring GPP at Kiona and five environmental factors were reviewed to determine the relative importance of each factor: (1) the daily mean streamflow at Kiona, (2) the daily median turbidity measured at Kiona, (3) the photoperiod (time between sunrise and sunset), (4) the daily mean photosynthetically active radiation (PAR) measured at a nearby meteorological station (Prosser, WA) (Washington State University, 2007), and (5) the daily maximum water temperature measured at Kiona.

The strongest factor affecting spring GPP was daily mean streamflow, which was negatively related to GPP (fig. 28): greater values for GPP related significantly to lower values for daily mean streamflow. Figure 28 shows how the GPP value was more responsive to small increases in streamflow during periods of lower streamflow than during periods of higher streamflow. An increase from 500 to 1,500 ft³/s in the daily mean streamflow corresponded to a decrease of 6.64 g/m²/d
in GPP, whereas an increase from 4,000 to 5,000 ft³/s in the daily mean streamflow corresponded to a decrease of only 1.64 g/m²/d in GPP.

Although high streamflows can suppress aquatic plant growth by disrupting the streambed in which macrophytes are rooted and on which periphyton grow, no known bed-disturbing events occurred in the Kiona reach during this study. Therefore, the decrease in spring GPP that corresponded to high streamflow likely was due to the increased turbidity and water depth (and thus a decrease in light availability) associated with the high streamflow. The effect of water depth on GPP could not be determined because it was not independent from the GPP calculation, but the effect of turbidity on GPP could be, given that GPP was related negatively to the daily median turbidity (fig. 28). The
mean GPP when the turbidity was below 10 FNU was
12.95 g/m²/d and was significantly greater than the mean
value of 3.11 g/m²/d when the turbidity was above 10 FNU.

Spring GPP was related positively (but weakly) to the photoperiod and the daily maximum water temperature, but during periods of low streamflow (and greater light availability) GPP also was related to daily mean PAR (table 14). All these variables are expected to influence plant growth in the Yakima River, but none can be considered in isolation because of the interdependence between them.

Evaluation of the Gross Primary
Productivity Method

The results from this study showed that spring GPP estimates for the Kiona reach were consistent with the maximum macrophyte biomass for a growing season and provided information on the important factors affecting macrophyte growth during this critical period in the life cycle of the plants. In contrast, the summer GPP estimates for the Kiona reach were not consistent with the maximum macrophyte biomass for a growing season.

Other researchers have obtained GPP estimates that did not appear to correspond to the level of macrophyte biomass. Thyssen (1982) speculated that increased metabolic activity from benthic algae was the reason that he determined no significant effect on photosynthesis and respiration in a stream when macrophytes were removed. Kaenal and others (2000) determined that macrophytes and their associated epiphytes contributed to stream metabolism in nutrient-rich, unshaded streams, that stream metabolism may not be dominated by macrophytes even when macrophyte biomass is high, and that GPP recovered quickly after macrophyte removal probably because of the growth of benthic algae where plants previously had grown.

Although this study did not attempt to determine the relative contribution of different plant types (periphyton, epiphyton, and macrophytes) to the total plant productivity, the results strongly indicated that aquatic plants other than macrophytes were responsible for much of the productivity occurring in the Kiona reach during summer. Abundant epiphytic algal growth commonly was observed on the macrophytes in the Kiona reach, filamentous green algae was observed growing in open spaces between plant beds, and the mean rate of periphytic algal accrual measured in the reach during one periphytometer experiment in 2007 was greater than most of the mean accrual rates measured in the Zillah reach in 2006 and 2007 when periphyton was at nuisance levels in some areas.

The GPP methodology is not appropriate when large oxygen loss from the water occurs in the form of bubbles (Britton and Greeson, 1989), when oxygen transfer to the atmosphere occurs through emergent plants, or if heterotrophic respiration is a large part of the community respiration (Bales and Nardi, 2007). During summer 2005 bubble formation in and around the densest macrophyte beds in the Kiona reach was common and macrophytes were emergent in large areas of the reach. These plants often had numerous heterotrophic organisms attached such as Simulid black fly larvae, Brachycentrus caddisflies, and other types of organisms. Although all these factors might have caused GPP to be underestimated during summer 2005, another indicator of primary productivity, the median daily range in pH, strongly indicated that summer GPP in the Kiona reach was similar between 2005–07. The median daily range in pH between July 1 and September 30 was not significantly greater in 2005 compared to the median daily ranges in 2006 and 2007, even though there was substantially more macrophyte biomass in 2005.

A comparison of the one-station and two-station whole-stream productivity results showed that, although the one-station method is adequate in reaches with low levels of macrophyte growth, the two-station method might be necessary in reaches with abundant macrophyte growth to account for variations in reach conditions that cannot be represented by one water-quality station. The differences between the one-station and two-station estimates at Zillah during summer 2005 and at Kiona during summer 2007 were generally low (< 11 percent); however, all but one of the two-station estimates at Kiona during summer 2005 were greater than the one-station estimates, and the difference between the two estimates averaged 17 percent and was as high as 62 percent.