Changes in nutrient ratios drive changes in pelagic and benthic assemblages, and benthic-pelagic coupling in Puget Sound: A compelling hypothesis linking water quality and the benthos.

Christopher Krembs1, Maggie Dutch1, Valerie Partridge1, Sandy Weakland1, Julia Bos1, Skip Albertson 1, Brandon Sackmann2, Mya Keyzers1, Laura Friedenberg1, Carol Falkenhayn Maloy1 1 Environmental Assessment Program, Washington State Department of Ecology, 300 Desmond Dr., Lacey, Washington, 98504, USA. E-mail: [email protected] Publication No. 14-03-024 2 Brandon Sackmann, Integral Consulting 1205 West Bay Drive NW Olympia, WA 98502, USA. E-mail: [email protected]

Fig. 9 Hypothesis:Hypothesis: ChangesChanges inin thethe MarineMarine FoodFood WebWeb andand EnergyEnergy TransferTransfer inin PugetPuget SoundSound

Introduction Drawing by: Christopher Krembs Results and Discussion

Microbial-based food web Noctiluca Zooplankton Fish Analyses of Ecology’s long-term monitoring data indicate Long-term increases in nitrogen concentrations (Fig. 1 A) and shifting nutrient ratios (Fig. 1 B) that changes in the cycling of organic material might + Nitrogen 10% 10% 10% suggest human nitrogen inputs to Puget Sound. Yet, decreasing phytoplankton biomass (Fig. 2) in occur that could affect phytoplankton, micro- our monitoring network (Fig. 3) suggests large-scale changes in lower trophic levels of the pelagic Fig. 3. Marine long-term zooplankton, and benthic communities. + nutrient cycling in the water monitoring locations in the Diatom-based food web Changing food web and more near-surface nutrient cycling food web that match a decline in the marine benthos (Fig. 4 A, B). greater Puget Sound area (southern Salish Sea). Si:N The long-term change has potential implications for Red dots depict monthly- • Monitoring of lower food web dynamics remains a knowledge gap in Puget Sound and is visited water-column 10% marine food web structure, energy transfer, particle 10% indispensable to connect water quality to the marine food web! stations. Colored regions export, and higher trophic levels such as fish. A testable illustrate benthic monitoring regions sampled using a hypothesis is presented combining observations about Human pollution causes nutrient imbalances (Fig. 1 B) and plays a role in the shift from diatom to dinoflagellate assemblages [1]. Large random sampling design. the significance of energy flow through the microbial Less sinking blooms of the dinoflagellate Noctiluca (Fig. 5) in our station network [2] are consistent with the increases in nitrogen (Fig. 1A). In contrast, a

Nitrogen of diatom Decreased coupling between the water and sediment decrease in subsurface phytoplankton biomass (Fig. 6), particularly in the late summer (Fig. 7), contradicts nitrogen trends. food web that determines the outcome of energy particles available for higher trophic levels. -- nutrientnutrient cyclingcycling inin thethe sedimentsediment • Could long-term observations indicate that food web dynamics alter nitrogen concentrations BenthicBenthic animalsanimals DecliningDeclining communitycommunity ofof organismsorganisms inin thethe sedimentsediment in the water?

Water Column A negative correlation between nitrate and phytoplankton biomass* implies that phytoplankton can control nitrate levels in Puget Sound. Therefore, mechanisms that control phytoplankton biomass could increase nitrogen concentrations. Hypothesis A. 2.5 • We present several observations that could indicate changes in the microbial food web in 2.0 y = 0.2007x - 402.49 Nitrogen additions to Puget Sound cause nutrient ratios Benthos R² = 0.4003 30m) 1.5 Si:N:P to change and indirectly might create a larger Puget Sound. 1.0 0.5 energy transfer through the microbial food web by Micro-zooplankton grazers (<200µm) control phytoplankton biomass in Dabob Bay, Hood Canal [3] and can consume the diatom biomass 0.0 promoting conditions for increased micro-zooplankton once per day [4]. We confirm that phytoplankton biomass and ammonium track one another following blooms (Fig. 7). We propose that -0.5 grazing. This has consequences for overall findings in Dabob Bay might apply to the larger Puget Sound region. A. Total Abundance Taxa Richness -1.0 phytoplankton species composition, biogeochemical 20% -1.5 10% • A shift from a copepod- to a micro-zooplankton-dominated food web can reduce the energy n a -2.0 cycles, higher trophic levels food availability, and i d Nitrate anomalies ( Nitrate µM, 0 - flow to higher trophic levels (more trophic levels) and reduce carbon export to depth. e 0%

-2.5 M benthic-pelagic coupling. n

i -10% 1998 2000 2002 2004 2006 2008 2010 2012 2014 e g

Abundant flagellate blooms in Puget Sound [2] (Fig. 5) and Noctiluca have been implicated in coastal eutrophication [5]. Noctiluca feeds on n -20% a h *

diatoms, copepod eggs, fecal pellets [5]. C

-30% t

25 n B. e *

c -40% EOPS, Aerial photography 6-17-2013 r • The export of organic material from a micro-grazer-dominated food web retains nutrients e 20 P *

-50% y = -0.9223x + 1855.4 near the surface [6]. It differs from the organic-carbon-rich, and fast-sinking fecal pellets of a * = significant 15 -60% *

30m) R² = 0.3423 - copepod-dominated food web. * * 10 Elliott Bay Elliott Bay Hood Canal Hood Canal San Juan Is. San Juan Is.

5 Evidence of a potentially reduced carbon export to depth comes from Ecology’s long-term benthic monitoring program. Puget Sound-wide South Sound South Sound Central Puget Central Puget All 6All Regions 6All Regions Str. of Georgia Str. Str. of Georgia Str. Whidbey Basin Whidbey Basin Whidbey Commence. Bay Commence. Bay Commence. Bainbridge Basin Bainbridge decreases in benthic endo- and epi-benthic detritivores can’t be tied to pollutants (Fig. 4). Basin Bainbridge All 3 Urban Bays 3 Urban All All 3 Urban Bays 3 Urban All 0 Si:N Si:N anomalies (0 sail boat -5 • We suggest a potential connection between the changes in the pelagic and benthic food ) 2 B. - -10 webs. 1998 2000 2002 2004 2006 2008 2010 2012 2014 We hypothesize a transition from diatom- towards a flagellates-dominated microbial food web, caused by nutrient imbalances (Fig. 9). Fig. 1. Inter-annual changes in (A) nitrate and (B) the ratio of silicate to dissolved Noctiluca inorganic nitrogen (Si:DIN) appear to be a result of disproportional nitrogen additions could reinforce the process by exerting grazing pressure on diatoms while competing with copepods for resources. to the system. From 1999 to 2013, inter-annual patterns of nitrate (A) steadily increased in the surface waters. As a result, the ratio of silicate to dissolved inorganic nitrogen (Si:DIN) decreased. Yearly anomalies were calculated by the difference between • Reduced phytoplankton biomass and strong micro-zooplankton grazing could explain the expected (site-specific seasonal baseline) and observed median water-column nitrogen increase via increased top-down grazing control on phytoplankton biomass. concentrations at 0, 10, and 30m. Anomalies from 17 stations were averaged over the year and station network (17 core stations) (see map Fig. 3). Fig. 5. Noctiluca bloom at surface in very long bands. Location: Between Bainbridge Island and Elliott Bay (Central Basin), 5:27 PM. Department of Ecology, Publication No. 13-03-075. * (data from Fig. 1A, correlated with Fig. 2, significant negative correlation, Spearman rank corr. coef.) Mean abundance no. of organisms 0.1 m

100 Fig. 4.

)

2 250 Declining 2.5 -2 - (A) Changes in benthic macro-invertebrate total abundance and taxa richness (both 0.1 m )

50m) 600

- )

Spring Max Summer Max Yearly AVG 2 80 - of 6 geographic regions and 3 urban bays of Puget Sound sampled approximately a decade , 0 2 m - Fig. 7. Seasonal patterns of subsurface phytoplankton, apart in time. Asterisks indicate statistically significant changes (Kruskal-Wallis test, α = 0.05).

500 a m

60 y = -5.5987x + 11251 ammonium, and Noctiluca. (B) Functional feeding guild [7] composition of benthic assemblages in 8 geographic regions a 200 2.0

)

R² = 0.5312 y = -15.896x + 32153 1 - Seasonal patterns of chlorophyll a (green line) and (left side of graphic) and 6 urban bays (right). Paired baseline and resample results are given 40 400 R² = 0.4633 ammonium (red line) concentration across Puget Sound for the 6 resampled regions and 3 resampled urban bays. Sample sizes are approximately 40 for give insight into the seasonal coupling of the microbial each region and 30 for each urban bay. 300 150 1.5 20 50m (mg Chl food web. Both chlorophyll a and ammonium are proxies average (mg Chl a m average - for algal standing stock and ammonium excretion by 0 phytoplankton grazers. Two seasonal phytoplankton

wide wide 200 100 1.0 peaks (spring and late-summer bloom) are followed by peaks in ammonium. Seasonal data are based on 10-year -20 100

- Sound y = -6.9338x + 14045 Puget Sound-wide averages over the period 1999 to Noctiluca blooms References: R² = 0.58 Ammonium conc. (µM l 2008. Data were spatially (0, 10, 30m depth) aggregated -40 50 0.5 0 2011 at 17 monthly-sampled stations. [1] Dongyan Liu et al., 2013 . 1998 2000 2002 2004 2006 2008 2010 2012 2014 Puget Chlorophyll anomalies (mg anomalies (mg Chlorophyll Chl 1998 2000 2002 2004 2006 2008 2010 2012 2014 2012 [2] Eyes Over Puget Sound (EOPS), 2011-2013. Chlorophyll content 0 2013 [3] Leising et al., 2005. Fig. 2. Declining phytoplankton biomass in Puget Sound. Fig. 6. Changes in the relative importance of spring and late-summer phytoplankton 0 0.0 [4] Verity et al., 1996a; Nejstgaard et al., 1997; Landry et al., 2000a; Calbet, 2001; Suzuki et al., 2002. The inter-annual decline in subsurface phytoplankton biomass suggests a reduced supply blooms. Long-term trends in chlorophyll a standing stock (0-50m) are decreasing (green). [5] Dodge, 1982; Fukuyo et al., 1990; Hallegraeff, 1991; Taylor et al., 1995; Steidinger & Tangen, 1996. 1 2 3 4 5 6 7 8 9 10 11 12 [6] Kiørboe, 2003. in organic material for the marine food web. The trend is based on site and seasonal The decrease is predominantly driven by a decline in the late-summer phytoplankton bloom Month [7] Macdonald et al., 2012. anomalies and calculated as in Fig. 1. (red), yet spring blooms are unchanged (black). Data were calculated as in Fig. 1.