Lagrangian-based studies in the coastal Gulf of Maine

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

Problem: LEO-derived productivity estimates presently rely on single images of stocks and state variables to infer rates of change

Solution: Multiple views/day from GeoCAPE will 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 in the Gulf of Maine

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

Part 2: Preliminary results of tracking particle inventories with a high resolution circulation model and MODIS 250m data

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

These are equivalent (Within the context of a homogenous water mass)

With GeoCAPE, we can probably track dPOC : dt

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

Time-depth biological Oxygen Anomaly (μmol/m3)

• Raw oxygen and f-chl profiles were corrected with bottle data

• bbp and c-660 derived particle inventories estimated using a regression with bottle POC

• Integration to euphotic depth

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

Data ProcessingData Processing

y = 0.1882x – 0.2243

r2 = 0.45

Results: optically-derived particle inventories versus NCP

r2 = 0.85y = 0.1432x + 0.0124

Results: optically-derived particle inventories versus NCP

r2 = 0.76

y = 0.1058x – 0.2125

Results: optically-derived particle inventories versus NCP

• The Good: Highly significant relationships between optically derived particle inventories and NCP

• The Bad: The relationships (in carbon units) should be 1:1, but are off by a factor 3 - 5

• The Ugly: We don’t know why (yet!)

Conclusions for part 1Conclusions for part 1

Part 2: Preliminary results of tracking particle inventories with a high resolution circulation model and MODIS 500m data

Based on recent work: Estimating community productivity by tracking particle inventories in a Lagrangian context

Jonsson, Salisbury, Mahadevan, Campbell (2009) Jonsson, Salisbury, Mahadevan (2011)

POCt1

POCt2

Premise:

(POCt2 - POCt1)

(t2 - t1)NCP

POC inventory gives an estimate of NCP

Still to do on the GEO-CAPE Grant:

Use a 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 (12:27 AM last night)

The high res domain, Casco Bay Maine (about 120x120 km)

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)

mg

C m

-2 d

-1

Net community productivity (gC m2 d-1)

Interpolation of a MODIS chl row over 5 days

Linear

Tim

e (5

days

)

Longitude

Lagrangian

Many assumptions but the biggest are:1. Within the euphotic zone, along a Largrangian trajectory

POC = pCO2(bio)

2. Phytoplankton POC : Chl = 53

3. Sinking, vertical mixing and DOC production by phytoplankton “excess production” are minimal, over short (2-7 day) time scales