4.5 Marine Food Webs
Energy Flow Through the Ocean
Marine food webs describe the complex network of feeding relationships that transfer energy from primary producers through consumers to top predators and decomposers. Unlike simple food chains, real marine ecosystems feature intricate webs with omnivory, size-based predation, and the microbial loop operating in parallel with the classical food chain.
The biological pumpβthe suite of processes that export organic carbon from the surface to the deep oceanβis intimately linked to food web structure. The efficiency of carbon transfer through trophic levels, the sinking flux of particulate organic carbon (POC), and mesopelagic carbon cycling together determine how much atmospheric COβ is sequestered in the deep ocean on timescales of centuries to millennia.
Trophic Levels & Transfer Efficiency
Energy is lost at each trophic transfer through respiration, excretion, and egestion. The trophic transfer efficiency (TTE) is typically ~10% (Lindeman's 10% rule):
$$\text{TTE} = \frac{P_{n+1}}{P_n} \approx 0.10$$
where $P_n$ is the production at trophic level $n$. After $k$ transfers: $P_k = P_0 \cdot \text{TTE}^k$
TL 1
Phytoplankton
50 Gt C/yr
TL 2
Zooplankton
5 Gt C/yr
TL 3
Small fish
0.5 Gt C/yr
TL 4
Large predators
0.05 Gt C/yr
Ecological Stoichiometry
The elemental composition (C:N:P ratios) changes across trophic levels. Zooplankton have lower C:N than phytoplankton, creating stoichiometric mismatches. Consumer-driven nutrient recycling (CNR) preferentially regenerates the element in least demand, shaping nutrient availability for primary producers.
Derivation: Trophic Transfer Efficiency
Step 1: Energy budget of a consumer
Consider a consumer at trophic level $n+1$ that ingests food at rate $I_n$ (energy per unit time from trophic level $n$). The ingested energy is partitioned into assimilated energy and egestion (feces):
$$I_n = A + F_{\text{eg}}$$
where $A$ is assimilated energy and $F_{\text{eg}}$ is egested energy. The assimilation efficiency is $\text{AE} = A / I_n \approx 0.6\text{-}0.9$.
Step 2: Partition assimilated energy
The assimilated energy is further divided into respiration $R$ (metabolic costs), excretion $U$ (dissolved losses), and production $P_{n+1}$ (growth + reproduction):
$$A = R + U + P_{n+1}$$
Step 3: Define trophic transfer efficiency
The trophic transfer efficiency (TTE) is the ratio of production at level $n+1$ to production at level $n$. It can be decomposed into three component efficiencies:
$$\text{TTE} = \frac{P_{n+1}}{P_n} = \underbrace{\frac{I_n}{P_n}}_{\text{exploitation}} \times \underbrace{\frac{A}{I_n}}_{\text{assimilation}} \times \underbrace{\frac{P_{n+1}}{A}}_{\text{net growth}}$$
Step 4: Estimate typical values
Each component typically contributes a fraction less than 1. For a marine zooplankton grazer: exploitation efficiency $\approx 0.4$, assimilation efficiency $\approx 0.7$, net growth efficiency $\approx 0.35$:
$$\text{TTE} \approx 0.4 \times 0.7 \times 0.35 \approx 0.10$$
This yields the classic "10% rule" -- only about 10% of energy at one trophic level is converted to production at the next level.
Step 5: Production available at trophic level $k$
After $k$ trophic transfers from a base of primary production $P_0$, the production available is:
$$P_k = P_0 \cdot \text{TTE}^k = P_0 \cdot (0.10)^k$$
For example, from 50 Gt C/yr of marine primary production: tuna at TL 4 receive $50 \times 10^{-3} = 0.05$ Gt C/yr. This explains why shorter food chains (e.g., sardines at TL 2-3) support far greater fishery yields than long chains (e.g., tuna at TL 4-5).
Derivation: Lindeman Efficiency and the Thermodynamic Constraint
Step 1: Lindeman's progressive efficiency
Lindeman (1942) defined the efficiency at each trophic level as the ratio of energy intake at level $n$ to energy intake at level $n-1$:
$$\lambda_n = \frac{\Lambda_n}{\Lambda_{n-1}}$$
where $\Lambda_n$ is the total energy assimilated at trophic level $n$.
Step 2: Thermodynamic constraint (Second Law)
The Second Law of Thermodynamics requires that entropy increase with each energy transformation. Not all assimilated energy can be converted to biomass -- a substantial fraction must be dissipated as heat through respiration:
$$\Lambda_n = R_n + P_n + U_n, \quad \text{with } R_n > 0 \quad \Longrightarrow \quad \lambda_n < 1$$
Step 3: Relate Lindeman efficiency to food web length
The total energy reaching trophic level $N$ from the primary production base $\Lambda_1$ is:
$$\Lambda_N = \Lambda_1 \prod_{n=2}^{N} \lambda_n$$
Step 4: Maximum food chain length
A trophic level can only sustain a viable population if $\Lambda_N$ exceeds some minimum threshold $\Lambda_{\min}$. This sets a maximum chain length:
$$N_{\max} = 1 + \frac{\ln(\Lambda_1 / \Lambda_{\min})}{\ln(1/\bar{\lambda})}$$
where $\bar{\lambda}$ is the geometric mean Lindeman efficiency. With $\bar{\lambda} \approx 0.10$, each additional trophic level requires 10x more primary production, limiting most marine food chains to 4-5 levels.
Step 5: Lindeman's observation of increasing efficiency
Lindeman noted that $\lambda_n$ tends to increase with trophic level (e.g., $\lambda_2 \approx 0.05$, $\lambda_3 \approx 0.15$, $\lambda_4 \approx 0.20$). This arises because higher predators are more mobile and better at locating and consuming prey, leading to higher exploitation efficiency. The overall pattern is bounded by:
$$\lambda_n \leq \text{AE}_n \times \text{NGE}_n \leq 0.9 \times 0.5 = 0.45$$
where AE is assimilation efficiency (bounded by digestive capacity) and NGE is net growth efficiency (bounded by metabolic costs). In practice, marine Lindeman efficiencies range from 5% to 25%.
Step 6: Implications for fisheries yield
Ryther (1969) used Lindeman efficiency to estimate potential fisheries production. Starting from net primary production and applying ecosystem-specific chain lengths and efficiencies:
$$Y = \text{NPP} \times e \times \prod_{n=2}^{N} \lambda_n$$
where $e$ is the export ratio and $Y$ is fisheries yield. Upwelling systems (short chains, $N = 3$, high $\lambda$) produce ~50% of world fish catch from <1% of ocean area, while open-ocean gyres (long chains, $N = 5$, low $\lambda$) are comparatively unproductive.
The Microbial Loop
Pomeroy (1974) and Azam et al. (1983) revealed that a large fraction of primary production (~50%) passes through dissolved organic matter (DOM) and is consumed by heterotrophic bacteria. This "microbial loop" recovers energy that would otherwise be lost:
$$\text{DOM} \xrightarrow{\text{bacteria}} \text{Flagellates} \xrightarrow{\text{grazing}} \text{Ciliates} \xrightarrow{\text{grazing}} \text{Mesozooplankton}$$
Classical Food Chain
Diatoms β copepods β fish. Short, efficient chain. Dominates in productive, nutrient-rich waters (upwelling, spring blooms). High export efficiency.
Microbial Food Web
Picoplankton β DOM β bacteria β flagellates β ciliates. Many trophic steps, more respiration, less export. Dominates in oligotrophic systems. Most carbon recycled in the surface layer.
The Biological Pump & Martin Curve
The biological pump transfers organic carbon from the surface to the deep ocean through sinking particles, vertical migration, and physical mixing of dissolved organic matter. The attenuation of sinking POC flux with depth follows the Martin curve (Martin et al., 1987):
$$F(z) = F_{100} \left(\frac{z}{100}\right)^{-b}$$
$F_{100}$ = flux at 100 m (export production), $z$ = depth (m),$b \approx 0.858$ (global mean). Only ~1% of surface production reaches 4000 m.
~10 Gt C/yr
Export at 100 m
~2 Gt C/yr
Flux at 1000 m
~0.1 Gt C/yr
Reaches the seafloor
Carbon Transfer Efficiency
The transfer efficiency $T_{\text{eff}} = F(z_2)/F(z_1)$ varies with ecosystem structure: diatom-dominated systems have higher $T_{\text{eff}}$ (large, fast-sinking particles) while picoplankton-dominated systems have lower $T_{\text{eff}}$ (more recycling in the mesopelagic). The exponent $b$ varies from 0.5 (efficient) to 1.5 (inefficient).
Mesopelagic Carbon Cycling
The mesopelagic zone (200-1000 m) is where most sinking POC is remineralized. This "twilight zone" contains ~10 Gt C of fish biomass and hosts intense microbial activity. The balance between particle sinking and consumption determines the depth at which carbon is returned to dissolved form:
$$\frac{dF}{dz} = -\frac{F(z)}{L_r} = -\frac{r}{w_s} F(z)$$
$L_r = w_s / r$ = remineralization length scale, $w_s$ = sinking speed,$r$ = remineralization rate
Fast-sinking particles (fecal pellets, aggregates) penetrate deeper before remineralization. Fragmentation by zooplankton and microbial colonization reduce sinking speed, creating a complex interplay between physics and biology in the mesopelagic.
Export Production & the e-Ratio
Export production is the fraction of primary production that sinks below the euphotic zone. The export ratio (e-ratio) varies with ecosystem structure and productivity:
$$e = \frac{F_{\text{export}}}{\text{NPP}} = \frac{F_{100}}{\text{NPP}}$$
Oligotrophic gyres: $e \approx 0.05$. Upwelling: $e \approx 0.20$. Spring blooms: $e \approx 0.30\text{-}0.50$.
The e-ratio increases with total production (Laws et al., 2000) and is higher in diatom-dominated systems (large, ballasted cells) than in picoplankton-dominated systems (small cells recycled in the microbial loop). Temperature also affects the e-ratio: cold waters export more efficiently due to reduced microbial respiration rates in the mesopelagic.
Python: Martin Curve, Trophic Cascade & Biological Pump
Python: Martin Curve, Trophic Cascade & Biological Pump
Python!/usr/bin/env python3
Click Run to execute the Python code
Code will be executed with Python 3 on the server
Fortran: Size-Structured Ecosystem Model
Fortran: Size-Structured Ecosystem Model
Fortran-------------------------------------------------------
Click Run to execute the Fortran code
Code will be compiled with gfortran and executed on the server
Key Takeaways
- βΈTrophic transfer efficiency is ~10%; most energy is lost as respiration at each level.
- βΈThe microbial loop processes ~50% of primary production through bacteria and protists.
- βΈThe Martin curve $F(z) = F_{100}(z/100)^{-b}$ describes POC flux attenuation with depth.
- βΈExport production (~10 Gt C/yr at 100 m) drives deep ocean carbon storage; only ~1% reaches the seafloor.
- βΈFood web structure determines the biological pump efficiency: diatom chains export more than microbial webs.