Barr et al Ecosummit presentation Oct 2012, Columbus OH

Remote sensing of mangrove forest productivity
Ecosummit, Columbus OH, Oct 2012

1. Effects of natural disturbances, salinity and temperature on light-use efficiency and carbon cycling in the subtropical mangrove forests of Everglades National Park Jordan G. Barr 1 , Vic Engel 2 , Jose D. Fuentes 3 , Douglas O. Fuller 4 1 South Florida Natural Resource Center, Everglades National Park, Homestead FL 33030 2 Southeast Ecological Science Center, U.S. Geological Survey, Gainesville FL 32653 3 Department of Meteorology, Pennsylvania State University, University Park PA 16802 4 Department of Geography and Regional Studies, University of Miami, Coral Gables, FL 33124

2. Outline - LUE in mangrove forests Mangrove forest characteristics Coastal carbon budgets and definitions Partitioning NEE into GPP and RE Modeling light-use efficiency (LUE) and gross primary productivity (GPP) Disturbances result in spatiotemporal changes mangrove forest productivity Conclusions

3. Everglades National Park Riverine mangrove forest dominated by R. mangle, A. germinans , L. racemosa Mangrove forest structure varies across the salinity mixing zone Growth and total biomass reflect [P] gradients and peat depths (up to 6 m) Semi-diurnal tides Distinct wet and dry seasons Avg. precipitation: 160 cm yr -1 Avg. ann. air temperature: 24 C

4. Everglades National Park Study site Height (m) b 19.0 ± 0.2 Basal area (m 2 ha -1 ) a 39.7 Tree density ( ind. ha -1 ) b 7450 Flood frequency (floods yr -1 ) b ~ 500 Flood d uration (h yr -1 ) b 4622 (52 % of time) Total: N (%) a 1.2 ± 0.1 P (%) a 0.12 ± 0.01 C (%) a 22.2 ± 1.2 Bulk density (g cm -3 ) b 0.212 a Chen and Twilley , 1999 b Krauss et al. 2006

5. Net ecosystem carbon balance (NECB) in riverine mangrove forests NECB = -  NEE + F DIC,DOC,POC (Chapin et al. 2006) NEE == Net ecosystem exchange of CO 2 NECB == Net ecosystem carbon balance

6. Ecosystem respiration R E response to temperature mirrors that of leaf-level carboxylation and ecosystem gross primary productivity (GPP).     R E declines at higher temperatures Response functions used to estimate daytime R E to derive GPP.

7. Seasonal C exchange patterns Daily GPP values reach a plateau during March to November. Lower wintertime GPP values are due to reduced temperature and day length. Daily LUE values are maximal during September to November as a result of: lower PAR (30-50 mol photons m -2 day -1 ) compared to levels in April to June, near optimal (27 C) air temperature, and reduced salinity levels (10-20 ppt ). Daily PAR values are maximal in May before the onset of the wet season and are minimal during the winter solstice.

8. Modeling light use efficiency (LUE) EVI from MODIS m EVI = 4.03 ± 0.52 /       FAPAR Linear decline in efficiency with increasing PAR and salinity levels. Efficiency declines both above and below some optimal air temperature Fraction of PAR absorbed by photosynthetic active vegetation asymptotes to 1 with increasing EVI.

9. Modeling LUE – advantages of Bayesian framework The Bayesian modeling framework combines information about the data with any prior (expert) knowledge of the forcing terms. Compute uncertainties in modeled mean LUE and variance in LUE (not shown) Model output includes the distribution, and therefore the significance, of all forcing terms (i.e., parameters).

10. LUE model performance Our LUE model: 2004-2005 slope = 0.720, intercept = 0.144, R 2 = 0.56 2006-2011 slope = 0.483, intercept = 0.249, R 2 = 0.45 MODIS Our model

11. LUE model conclusions Mangrove forests exhibited a maximum light-use efficiency (quantum yield) of 3.2±0.2% . Effect of temperature: 1) Optimal LUE at 28 C , 2) LUE declines to zero for air temperatures < 3 C and > 35 C . Co-occurrence of high PAR (>50 mol PAR m -2 day -1 ) and high salinity (>30 ppt ) contributed to reduced LUE during May and June. 4.7% reduction in LUE per 10 ppt increase in salinity 10.1% reduction in PAR per 10 mol photons m -2 day -1 increase in PAR (photo-saturation in mangrove forests) Effect of EVI on FAPAR was significant, but contributed to a modest 20% to 30% seasonal variability in LUE.

12. Net ecosystem carbon balance (NECB) in riverine mangrove forests SR mangrove forests represent outliers in terms of forest-atmosphere C exchange with S -NEE of 956±175 g C m -2 per year Mean C burial rates are 124  32 gC m -2 per year (95% C.I ., Breithaupt et al., 2012 ). Over hundreds of years, NECB  C burial rates, and NECB represents only 13% of S -NEE . Therefore, most (87% of NEE) may be exported to the estuary as DIC, DOC, and POC .   - S NEE

13. Conclusions Mangrove forests in the Everglades sequester 956±175 g C m -2 per year. Much of this C ( 80% to 90%) may eventually be exported to the estuary and coastal ocean as organic (DOC, POC) and inorganic (DIC) carbon. Ecosystem respiration (RE) in mangrove forests obeys a temporally dynamic non-linear function of air temperature. Quantifying these patterns was essential for estimating seasonal patterns in GPP. Air temperature, daily PAR, salinity, and EVI represent significant controls on GPP. Our Bayesian modeling framework provided uncertainty estimates of both LUE and individual forcing terms. EVI values ( in the slides that follow ), determined from Landsat 5 TM, were both spatially and temporally variable during 1985 to the present. These trends will prove essential in understanding mangrove forest productivity in response to disturbances (i.e. hurricanes and cold periods), climate change, and upstream water management practices.

14. 1985.11.02 Spatiotemporal changes in EVI following disturbances (Red is high (~1), blue is 0)

15. 2004.01.23

16. 2005.11.09 Defoliation following Hurricane Wilma

17. 2006.03.01 Regrowth has not Yet occurred

18. 2009.10.19 Recovery of mangroves Adjacent to the coast

19. 2010.02.08 Mangrove forest transgression Jan 2010 cold period Destroys inland mangroves

20. 2011.11.10 Damage persists 2 years later

21. Acknowledgements J.D.F. and V.E. acknowledge support from the DOE (grant DE-FC02-06ER64298) to participate in this research. This research is also based upon work supported by the National Science Foundation through the Florida Coastal Everglades Long-Term Ecological Research program under grant number DBI-0620409.

22. Extra slides… Parameter Description Mean SD 2.50% Median 97.50% e Optimum light-use efficiency (mmol C (mol photons) -1 ) 31.8 2.2 27.7 31.6 36.7 m EVI Curvature of FAPAR response to EVI (dimensionless) 4.03 0.52 3.11 3.99 5.21 T min Temperature minimum (C) 2.6 0.6 1.4 2.7 3.7 T max Temperature maximum (C) 33.5 0.6 32.4 33.5 34.8 T opt Temperature optimum (C) 27.8 0.3 27.2 27.8 28.5 m sal Salinity forcing (dimensionless) 0.0047 0.0022 0.0000 0.0048 0.0084 m PAR PAR-saturation forcing 0.0101 0.0004 0.0092 0.0102 0.0110

23. Figures Annual GPP vs RE

24. The DEM's for the state of Florida were developed by staff at NOAA's Coastal Services Center, Charleston, SC. The DEMs were derived from a variety of lidar data, which can be obtained via the Digital Coast website (csc.noaa.com/ digitalcoast ). The data are broken down by NOAA National Weather Service Weather Forecast Office boundaries, as follows : Miami (MFL) MFL_1_GCS_NAVDm.img Collier, Monroe (excluding Keys ) The DEMs are provided in geographic coordinates (NAD83) and vertically referenced to NAVD88. Vertical units for all data are meters . Portions of some DEMs may have data voids due to the fact no lidar data was available for the area at the time of development. Check the Digital Coast for our latest lidar holdings. Any questions/concerns regarding the DEMs or derived products used in the Sea Level Rise and Coastal impacts Viewer, please contact : Doug Marcy, [email protected], 843-740-1334

25. Magenta: -0.45 m Blue: -0.1 m Cyan: +0.1 m Green: 0.65 m Yellow: 1.0 m Red: 1.4 m

26. EC flux tower