Lagrangian-based estimates of Net Community Production Overall goal: estimation of community production rates by tracking satellite-derived C inventories.

Lagrangian-based estimates of Net Community Production

Overall goal: estimation of community production rates by tracking satellite-derived C inventories over time and space

Problem: Various satellite productivity estimates presently rely on single images of stocks and state variables to infer rates of change. They do account for actual changes over time.

Potential solution: Multiple views/day from GeoCAPE or a swarm of LEO sensors can enable monitoring of biogeochemical inventories within a water parcel as they evolve over time.

Lagrangian-based Studies in the coastal Gulf of Maine (supplements to NASA-Carbon NNX08AL80G).

Part 1: A Lagrangian field experiment to determine net community productivity NCP in the Gulf of Maine

Olivia DeMeo (MS Student, UNH), Joe Salisbury (UNH)

Part 2: Tracking particle inventories with a high resolution circulation model to infer NCP

Bror Jonsson (Princeton), Joe Salisbury (UNH), Amala Mahadevan, (Boston University)

Part 1: The Lagrangian Experiment

Approach: 1. Tracked a drogue at 12m (7 cruises over 16 days)2. Kept track of oxygen and particle inventories

We assume these are equivalent (within the context of a homogenous water mass)

• Raw oxygen, bbp and c-660 and f-chl from profiles were corrected with bottle data

• Oxygen inventories corrected for thermodynamic variability, air-sea flux and diffusion, then converted to carbon using the Redfield ratio

• Inventories (down to 1% surface PAR) were differenced over time

• Additional correction for DOCex, POC flux and bubble injection are underway

Data/ data processingData/ data processing

Cruise data examples: Time-depth f-chlorophyll (mg/m3)

Time-depth biological Oxygen Anomaly (μmol/m3)

y = 0.1882x – 0.2243

r2 = 0.45

Results: optically-derived particle inventories versus NCPChlorophyll (C:Chl = 90:1)

r2 = 0.76y = 0.1058x – 0.2125

r2 = 0.85y = 0.1432x + 0.0124

Newly corrected for DOCex, POC flux and bubble injection. The slope is close to 1

• The Good: Significant relationships between optically derived particle inventories and NCP. Corrections for DOCex, POC sinking and bubble injection brought slopes close to one.

• The Bad: oxygen > particles. We either have too much oxygen or too few particles. Very dependant on the above corrections.

Conclusions for part 1Conclusions for part 1

Premise:

Part 2: Estimating NCP by tracking sat particle inventories in a Lagrangian contextJonsson, Salisbury, Mahadevan, Campbell (2009) Jonsson, Salisbury, Mahadevan (2011)

PCt1

PCt2

(PCt2 - PCt1)

(t2 - t1)NCP

PC = satChl * *ze

ze =loge(0.01)=K490

modified from Behrenfeld et al., 2005***Movie examples shown w/ chl

PC inventory helps give an estimate of NCP

chl

mg

C m

-2 d

-1

Net community productivity (gC m2 d-1)

What we’re working on now: Time and space requirements for satellite data to enable particle inventory tracking

300m, hourly model and daily cloud free 250 and 500m MODIS data to:

1.Simulate differences in “net radiance production” between Eularian versus Lagrangian determinations over the course of a day.

2.Run the same simulation using increasingly large pixel resolutions.

The first run using high res circulation and 500m MODIS

How do our results help inform the GEO-CAPE SWG?

1.Results from part 1 suggest that sub daily changes in particle inventories can be use to to track daytime NCP rates2.3-5 determinations per day may be enough for daily NCP estimates provided the advective component is adequately resolved3.For part 2: Preliminary work shows promise towards estimating rates from satellite tracking of particle inventories in a Lagrangian context.4.In work still to be done, we anticipate considerable differences between the Lagrangian and Eularian approach (using high resolution data)

Laws and Bannister (1980)

Behrenfeld et al., 2005

For which is substituted by

CARBON TO CHLOROPHYLL DETERMINATIONS

Interpolation of a MODIS chl row over 5 days

Linear

Tim

e (5

days

)

Longitude

Lagrangian