Tahoe Stormwater Assessment and Management for Urban …...This wetland was constructed specifically to treat stormwater runoff from the urban core of Tahoe City and was brought on-line

review (Bernal and Mitsch 2012) were also lower than in these created wetlands The rates of C burial in lakes and ponds of the conterminous US is a mean of 46 g C m-2 yr-1 and a median of 31 g C m-2 yr-1 which are much lower than found in this study (Clow et al 2015)

However rates of C burial in reservoirs where the sedimentation rate is much higher gave a mean of 149ndash363 g C m-2 yr-1 (Clow et al 2015) which was more comparable to the rates of C accretion reported in this article

Phosphorus assimilation of wetlands summarized from a large database (including both natural and treatment wetlands) suggested that the long-term assimilative capacity was 1 g P m-2 yr-1 although they could exceed this capacity in the early stages (Richardson et al 1996) The P accretion rate of both the Tahoe City wetland and the Olentangy wetlands (Mitsch et al 2014) exceeded this rate by almost 4 times in their first 15ndash16 yr after creation

Chemical fractionation and digestion would be necessary to distinguish the forms of P in the Tahoe City wetland but in suspended sediments of Ward Creek near the study site Ferguson (2006) found that 25 was bioavailable Of total P 67 was inorganic and 33 was organic In urban runoff from Tahoe City however 36 was bioavailable while 64 was inorganic P Thus statistical correlation between organic matter and P in the soil of the wetland basin is similar to that found by fractionation of the suspended sediments This distribution between organic P and inorganic P is typical for wetland soils (Qualls and Richardson 1996) Consequently our results lent support to hypothesis 3 but storage of organic P in the O horizon was still a significant portion of the accretion

The annual accretion rates for N and P from the wetland sediment cores correspond remarkably well with measurements of retention during a single year (calculated from inflow minus outflow concentration expressed as g m-2 yr-1) (Heyvaert et al 2006) The retention was 13 g total N m-2 yr-1 and 32 g total P m-2 yr-1 compared to 177 for N and 37 for P accretion in the cores (Figure 2) The slightly greater accretion of N could conceivably have been due to N fixation by cyanobacterial mats or simply difference in the loads over time The inflow concentrations to the Tahoe City wetland in 2003 tended to be similar to those from other commercial and highway areas of the Tahoe Basin but were relatively high compared to other watersheds with less urban land use (Heyvaert et al 2006) Also water year 2013 precipitation was about 90 of average over the 16-year period (NRCS SNOTEL site 809 in Tahoe City)

Sediment accumulation rates in urban stormwater ponds and constructed wetlands vary considerably More and Hunt (2012) reported accretion rates compiled from various studies ranging from 01 to 13 kg m-2 yr-1 High accretion rates can lead to a rapid rise in base elevation that ultimately limits the useful life of treatment systems Slow elevation rise would be considered something less than 125 cm yr-1 In the absence of site-specific accretion data the USDA Natural Resources Conservation Service recommends a minimum design standard of 254 cm yr-1 for constructed treatment basins (NRCS 2010) The Tahoe City wetland

18

accretion rate averaged 32 cm yr-1 Sedimentation (accretion) rates in the Olentangy wetlands (Mitsch et al 2014) ranged from 12 to 14 cm per year or 60 kg m-2 yr-1 which was comparable to the rate of mass accretion found in the Tahoe City wetland (769 kg m-2 yr-1) although the rate of elevation rise was greater in the Tahoe City wetland Mitsch et al noted that their average rate was very high compared to most wetlands because the wetlands were created in excavated basins undergoing ldquoprimary successionrdquo that were filling with sediments whose source was riverine water with substantial loads of sediment They noted that created wetlands generally have very high hydraulic loading rates high sediment loads and are newly created basins in contrast with natural wetlands where elevation rise is more in equilibrium with periodic erosion or decomposition The Tahoe City wetland fit into this pattern because it was also initially an excavated basin whose bottom was below the inflow and outflow elevations Another example of seven constructed wetland basins in Sweden specifically designed to trap sediment and phosphorus from water with high sediment and water loads of agricultural watersheds was summarized by Johannesson et al (2015) The sedimentation rates ranged from 13 to 108 kg m-2 yr-1 a range that encompassed the rate in the Tahoe City wetland but was far greater than most natural wetlands Rates of sedimentation in reservoirs of the conterminous US averaged 44 cm yr-1 (Clow et al 2015)

In the Tahoe City wetland the low spatial variability of Fe Mg and Mn concentrations that were associated with the inorganic sediments filling the basin suggested that sediment entered the wetland was evenly distributed and then settled Given the findings of Mitsch et al (2014) that rates of sedimentation on the order of 60 kg m-2 yr-1 are high and from our finding of association of metals with inorganic sediment accretion we conclude that rates of accretion of metals are also high in the Tahoe City wetland compared to most Thus measured rates of accretion of C N P and metals in the Tahoe City wetland supported hypothesis 1

One factor contributing to the rise in surface elevation that eventually impeded flow through the system was low bulk density of the unconsolidated soil material (Table 1) The bulk density in the Tahoe City wetland (Table 1) was only about 52 of that found in the sediments and accumulated organic matter in the Olentangy wetlands (Mitsch et al 2014) The buoyant force of the water in a continuously saturated wetland (or pond) as well as the content of low density organic matter contributed to low bulk density Wetlands or stormwater retention basins that experience periodic lowering of the water surface may experience more consolidation and less elevation rise but increased mineralization of C and N may be an associated cost

Watershed yield of total N (Table 3) was similar to that for an average for the State of California (Smith and Alexander 2000) Sediment yield however substantially exceeded the watershed yields from Tahoe Basin streams Rios et al (2014) summarized data from ten USGS monitored streams and compared these to the suspended sediment yield from Barton Creek near Tahoe City with values ranging from a low of 084 kg ha-1 yr-1 at Logan Creek to

19

a high of 163 kg ha-1 yr-1 at Third Creek In Barton Creek which is much less densely urbanized than Tahoe City the sediment yield was 816 kg ha-1 yr-1 compared to 1232 kg ha-1 yr-1 into the Tahoe City wetland This illustrates the tremendous load put onto relatively small treatment systems compared to larger drainage areas (290ndash14200 km2) with more natural conditions and less relative impervious area (05ndash79)

Despite the gleyed color of the soil Fe concentrations were similar to those of surrounding lowland soils of the watershed Wetland soils tend to lose Fe by reduction and leaching There may have been some reduction and leaching of Fe but the Fe content was still comparable to the surrounding soils at the time of sampling The export of Fe particularly reduced Fe is regarded as critical since Fe has been found to be limiting to algal growth of phytoplankton in Lake Tahoe (Goldman 1964) However the wetland was responsible for trapping a very large load of Fe in the accreted inorganic matter before it could be transported to Lake Tahoe or the Truckee River In this case the wetlandrsquos role as a stormwater retention basin which promoted the settling of sediments was probably the most important factor Mn is another redox reactive element but the concentrations were far lower than those of Fe (Table 2)

Pollution of Na from salt applications to roads has resulted in elevated concentrations in lakes and damage of vegetation in cold climates over the developed world Large quantities of Na are distributed on the roads of the Tahoe Basin each year where the average annual snowfall is 467 m (in Tahoe City) Since 22 of the watershed of the Tahoe City wetland is comprised of roads we might expect some excess salinity in the wetland However total Na is no higher than it is in the dominant soil type (the Tahoe series) Salt damage to roadside vegetation is common in the Lake Tahoe Basin (Fan et al 2014) There was evidence that Na was more concentrated in the O horizon closer to the surface (Table 2) which might indicate aerial deposition In fact Fan et al (2014) found that aerial spray from traffic and wind caused the greatest proportion of tree damage and was mostly confined to areas 10 m from roads

Boron is a plant micronutrient but it can be toxic to plants depending on the concentration Boron concentration can reach toxic levels in geothermal areas of the region (Smith et al 2013) but concentrations were relatively low in the wetland

ORIGINS OF MATERIAL ACCRETED IN THE WETLAND

Elemental composition is frequently used to infer the origin of sediments (eg Janneke et al 2012) Elemental composition in the mineral horizon correlated best with the B horizon of the Marla soil series compared to other soil series in the watershed that had geochemical data (Soil Survey Staff 2016 query ldquogeochemical datardquo) The correlation of concentrations of the elements listed in Table 2 vs the concentrations in the Marla soil geochemical data were significant with an R2 of 08 (Zn Cu and B were not included because they were not analyzed in the Soil Survey data) Concentrations of total Fe Mg and Mn Ca and K in the wetland mineral horizon were lower than in soils from the backslope

20

areas of the watershed (the Tahoma and Jorge series) This comparison suggests that most of the material came either from alluvial soils of the watershed or from traction sands with similar alluvial origins but not predominantly from erosion of the surrounding mountain slopes The concentration of C in the mineral horizon of the wetland was lower than in the A horizons of the soils of the watershed which also suggests that erosion of A horizon material was not the major source of sediment

ACCRETION OF ALLOCHTHONOUS VS AUTOCHTHONOUS ORGANIC NUTRIENT ELEMENTS

There is no simple way to measure the contributions of allochthonous C N and S to the accretion rate without a detailed budget of sediment in the inflow and outflow and net primary production over the 16 year period But with two assumptions we can make an estimate The mineral horizon had a minimum C concentration of 167 If we assume this value represents the C concentration in incoming sediment we can estimate the C accretion due to sediment inflow Assumption number two is that the organic matter originating from allochthonous source material contained 4 ash based on another study (Qualls and Richardson 2000) The additional C in the mineral horizon would then result primarily from root production In the O horizon there was also deposition of sediment as indicated by the concentration of inorganic matter but the balance of C would result from net primary production Using these two assumptions a mixture of sediment with 167 organic matter in the mineral horizon with 26 C would indicate that the difference of 093 C concentration resulted from root litter production in the mineral horizon In the O horizon we used the concentration of ash in the O horizon as a conservative tracer in equation 1 The contribution of sediment in the O horizon was estimated to be 50 of the mass but only 31 of the total C Thus 969 of the C in the O horizon was estimated to come from autochthonous C Using these percentages and the mass of C in each horizon (Tables 2 and 3) we estimated that 587 of the accretion of C in the whole soil profile had originated from autochthonous C The estimate was most sensitive to assumption 1 while the presumption of the ash content of decomposed litter was not critical (lt1 difference for a 50 difference in ash content) Consequently the contribution of the plant production in the wetland was the most important component of C accretion but sedimentation of allochthonous C was also a significant process

In the case of N the origins and end state are more complex because it is likely that most N entered the wetland as allochthonous inputs some of which were inorganic The very close association of C and N (R = 094) and the representative nature of the CN ratios (Mitsch and Gosselink 2015) for the end state of the wetland suggest that most N is bonded in association with the C resulting from primary production The rapid accumulation of the large O horizon resulting from net primary production required N to be taken up and deposited as organic matter that accumulated in the anaerobic conditions of the soil The fact that the O horizon was histic is evidence of inhibited re-mineralization of the cycled N If we do the same calculation (as outlined in the previous paragraph for the origins of C) for

21

nitrogen currently associated with the C originating from autochthonous production we estimate that 538 of the N is associated with autochthonous C This percentage is equivalent to an accretion rate of 11 g N m-2 yr-1 However the N entering the wetland had more diverse origins In the study of N inflow and outflow in 2003 (Heyvaert et al 2006) about 50 of the total N in the inflow was dissolved inorganic N (mostly nitrate) and about 59 of that was retained by the wetland The fate of this inorganic N could have included plant uptake and denitrification The estimate of total N retained by the wetland from the inflow minus outflow budget was 13 g N m-2 yr-1 a rate that is similar to the estimate of accretion currently associated with autochthonous C (95 g N m-2 yr-1) and the balance can be more than accounted for by N bound to C in the allochthonous sediment To account for the rapid accumulation of organic N associated with autochthonous C the demand exerted by uptake of vegetation alone could have accounted for most removal of N from inflow

Perhaps the most comparable rate of denitrification in the literature would be from the Olentangy wetlands where the denitrification rate was 27 g N m-2 yr-1 which accounted for only 2 of the N inflow (Batson et al 2010) In the Olentangy wetlands there was also evidence that plant uptake competed with denitrification for available nitrate That denitrification rate was also within the range of the other freshwater wetland studies reviewed in that study A review of N fixation rates in wetlands found them to mainly be less than 1 or 2 g N m-2 yr-1 but small scale measurements in cyanobacterial mats can be higher (Keddy 2010) In summary our findings support hypothesis 2 that a majority of C and N was eventually stored in association with plant production

EVIDENCE OF HIGHWAY RUNOFF

Since concentrations of major metals such as Fe and Mg may be similar in soils and highway runoff Zn and Cu have been used instead as tracers for highway runoff (Sriyaraj and Shutes 2001) Zn and Cu are products of tire and brake wear in particular and occur in highway runoff in elevated concentrations (Caltrans 2003) Concentrations in the O horizon of the Tahoe City wetland were much higher than in surrounding soils that have Zn and Cu analyses (Smith et al 2013) and are in a range termed ldquohighrdquo for levels of metal contamination in sediments as given by the Swedish Environmental Protection Agency (reported in Sriyaraj and Shutes 2001) The study of Sriyaraj and Shutes (2001) was particularly applicable since it concerned highway runoff from the M25 highway near London into a wetland and a retention pond Concentrations of Zn and Cu in the sediment were elevated compared to a control wetland They found that roots and rhizomes of Typha latifolia tended to concentrate Zn and Cu However the wetland and pond were effective in reducing the concentrations of heavy metals to acceptable concentrations in water flowing out Other studies have found that both Zn and Cu are taken up as micronutrients by wetland plants in greatly elevated concentrations when exposed to elevated concentrations in water or sediment (Shutes et al 1993 Weis and Weis 2004) and this effect is the basis for phytoremediation In fact the concentrations of Zn and Cu in the upper and lower depths of

22