Transcript OSM04_jushiro - IMECOCAL

IN SITU AND REMOTE SENSING ESTIMATED PRIMARY PRODUCTION IN THE
SOUTHERN CALIFORNIA CURRENT REGION, 1998-2002
350
J.C.A. CEPEDA-MORALES1 and G. GAXIOLA-CASTRO2
E-mail: [email protected]; [email protected]
ABSTRACT
In situ and modeled water-column primary production (PPeu) were determined from seasonally 1998-2002 IMECOCAL surveys and satellite data off Baja California. Behrenfeld and Falkowski (1997) model (VGPM) was applied for calculated PPeu, using PBopt obtained
for the southern California Current region from an empirical relationship between surface temperature (14.5°C-23.0°C) and in situ PB (Chl-a normalized rate of C fixation). We used SeaWIFS (Chl-a, EPAR, K490), and AVHRR (SST) imagery to feed the VGPM model. Regional PBopt has a
range from 1.1 to 8.9 mgC (mgChl–a h)-1, with an overall mean of 3.5 mgC (mgChl–a h)-1. PPeu estimated from the VGPM model with in situ data of Chl-a, PBopt and Zeu had a correlation coefficient of 0.88 when was related with in situ PPeu. However, when we used averaged monthly
satellite information, and the estimates of global and the regional parameters (PBopt), correlation coefficients diminished to 0.55 and 0.58 (p<0.05; n=175), respectively. When we use the global estimate of PBopt, PPeu was overestimated (slope of 1.28) in relation with in situ PPeu. Satellite Chla has the higher effect on lowering the correlation coefficient between in situ and modeled PPeu, with a difference of 0.37 in relation with PPeu modeled using in situ Chl-a. PBopt estimated from the B&F algorithm was one standard deviation higher than the overall regional mean. It is
necessary to take account these differences in the photosynthetic parameter when the VGPM model is used to estimate regional and local primary production from remote sensed information. In order to show the spatial variability of primary production from the VGPM model and the regional
PBopt, we separated the area in two main latitudinal regions divided by Line 113. In general, the southern region was most productive, with higher values during summer 1999-2002.
DATA AND METHODS
In situ primary production data were collected from 19 surveys realized
from 1998 to 2002 by the IMECOCAL program (Investigaciones
Mexicanas de la Corriente de California) off Baja California. In situ
PBopt values were obtained as the maximum Chl-a normalized rate of
14C fixation (PB = P/Chl-a) in the water column. Euphotic zone (Zeu)
integrated primary production (PPeu) was estimated from the in situ
experiments. Discrete values of Chl-a in the water column was
determined by the fluorometric method, from where surface
chlorophyll (Chl-asup) and (Chl-aopt) were obtained. From the CTD
profiles, temperature data (SST, Topt) were collected. Satellite
chlorophyll (Chl-asat) and K490 data were obtained from the SeaWiFS
program (DAAC, NASA). Monthly averaged EPAR were kindly gave by
Dr. R. Frouin from the Scripps Institution of Oceanography. Satellite
SST were obtained from the AVHRR-NOAA, PO.DAAC-NASA
program. AVHRR imagery are monthly composite data from January,
1998 to May, 2002, with a resolution of 9x9 km. A regional empirical
equation was calculated from PBopt and SST data. PPeu was calculated
using the Behrenfeld and Falkowski (1997) VGPM model, with
different estimates of PBopt.
Eo 
B 
PPeu  0.66125  Popt  
  Z eu  Copt  Di
 Eo  4.1
VGPM model (Behrenfeld y Falkowski, 1997)
To identify how the photosynthetic parameter and some variables affect
primary productivity estimations, we prove the effect on PPeu changing
PBopt, Zeu, and EPAR in the VGPM model. In the first scatter (A), we use in
situ data of Chl-aopt, Zeu, and global B&F PBopt. In this case PPeu was
overestimated, with a slope of 1.28, and a higher rms value. For the figure
B, global PBopt was changed by the regional estimator, and all the variables
were the same. The slope of the fitted line changed to 0.48, without a
significant difference in the linear correlation, but a smaller data
dispersion. In figure C, we use in situ values of Chl-aopt and PBopt, but
Zeu was changed by the values estimated from K490 satellite data. Linear
correlation coefficient increase, and the slope is close to one, and the data
dispersion diminished. For draw figure D, we calculated PPeu removing
the averaged monthly satellite EPAR from the VGPM model, keeping in
situ values of Chl-aopt, PBopt, and Zeu. We did not see any changes in the
correlation coefficient, either in the data dispersion, but with a better slope
.
Study area located at southern region of California Current off
Baja California. IMECOCAL grid survey program is shown
over a imagery of PPeu estimated for April, 1998.
RESULTS AND DISCUSSION
In situ water-column integrated PPeu values had a range from 8.0
mgC m-2 h-1 (October, 1998) to 240 mgC m-2 h-1 (October, 1999).
A higher percentage of the data (77.5%) had a range between 20
mgC m-2 h-1 to 80 mgC m-2 h-1, with a overall mean of 59 mgC m-2
h-1 (equivalent to 0.71 gC m-2 d-1).
In situ PBopt (mgC [mgChl-a h]-1) was showed together with the
parameters estimated with regional (red dots) and global (yellow dots)
algorithms (scatter A). Both empirical equations (B&F, and our
algorithm) have similar patterns, but with difference of the B&F
equation higher than one standard deviation (1.6) of the overall
regional mean (3.5 mgC [mgChl-a h]-1) (figure B). This pattern between
measured versus modeled PBopt has been also reported by Behrenfeld
et al. (2002) for Atlantic Ocean waters. The continuos line represent the
value of the regional ovearll mean, with dashed lines as one standard
deviation interval.
In situ PBopt values had a range from 1.0 mgC (mgChl-a h)-1 to 9.0
mgC (mgChl-a h)-1. About 76% of the data were between 2.5 and 5.0
mgC (mgChl-a h)-1, with an overall mean of 3.5 mgC (mgChl-a h)-1
(s.d.= 1.67). We had a wide dispersion of this parameter for each year,
with the tendency to increase during summer. This pattern was mainly
evident for 1999, 2000, and 2001.
Regional empirical relation between SST and PBopt was fitted by equation 1.
The full circles indicate the median of PBopt for each SST interval (0.5 oC).
The bars correspond to one standard deviation. Dashed contours show the
standard error of the estimations, and the solid lines is the best fit.
PBopt = -0.0012418 T 4 + 0.062177 T 3 - 0.93274 T 2 + 2.5485 T + 27.066
A
A Hovmuller diagram for each variable (SST,
EPAR, Chl-a, and Zeu) was done following a
latitudinal section parallel to the coast (100 km).
Figure A shows latitudinal changes of SST.
Figure B represent the temporal variability of
EPAR. Figure C describe the Chl-a patterns along
the transect, with a characteristic feature of the
Ensenada Front, and two main events with high
pigment concentrations during 1998, and 2002
years. Figure D show strong temporal and
latitudinal changes in the euphotic zone (Zeu).
Figure A shows the linear relationships between in situ PPeu and
modeled using monthy imagery composites of Chl-asat, and in situ
data of PBopt and Zeu. The linear correlation coefficient was <0.37
units, related with PPeu calculated with in situ surface Chl-a. When
we compare the in situ surface Chl-a with monthy composite
imagery (B), a very low correlation was found.
PPeu calculated from satellite data Chl-a, Zeu, and regional PBopt
estimated from the empirical algorithm and SST from satellite data is
shown in the scatter A. In figure B, PBopt was estimated by the B&F
global algorithm. In both cases linear correlation coefficients were low
(0.58, and 0.55 respectively) but statistically significant, with different
slopes.
C
A
(1)
B
Using remote sensed information (Chl-a, EPAR, Zeu)
monthly PPeu was calculated with the VGPM model and
regional PBopt algorithm. Figure A shows the overall
mean (1998-2002) of the PPeu (mgC m-2 d-1) estimated.
The line contours represent PPeu values along the study
area. Main features with high primary productivity (>400
mgC m-2 d-1) are the Ensenada Front and the inshore
waters. In the figure B, standard deviations of the overall
mean is shown.
D
Following Kahru and Mitchell (2000) we divided the area in four regions northern and southern of 28N. A first band from inshore to 100
km, and a second band from 100 km to 300 km offshore.
PPeu was modeled from the VGPM model with monthly imagery of
EPAR and in situ Chl-aopt, PBopt, and Zeu. Figure A shows a significant
correlation (r=0.88, p>0.05) between PPeu in situ vs PPeu modeled. In
the scatter B, Chl-aopt was replaced by surface in situ Chl-a. The
linear tendency is maintained but the correlation down 0.10 (r= 0.78,
p>0.05) and data dispersion increased, with a higher root mean square
(rms) value.
Temporal variability of monthly PPeu anomalies (monthly value minus overall mean) show higher values for the inshore southern (below
28°N) waters, without apparent differences at both offshore regions. During the El Niño 1997-98 were calculated the lower PPeu values,
with a more evident effect at inshore waters of the southern region. Higher anomalies were calculated during summer, with values >1.0 gC
m-2 d-1 for 1999-2002.
REFERENCES
Behrenfeld, M., and P. Falkowski. (1997). Photosynthetic rates derived from satellite-based chlorophyll concentration. Limnol. Oceanogr. 42. 1-20.
Behrenfeld, M., E. Marañon, D. Sieguel, and S. Hooker (2002). Photoacclimation and nutrient-based model of light-saturated photosynthesis for quantifying oceanic primary
production. Mar. Ecol. Prog. Ser. 228: 103-117.
Kahru, M., and G. Mitchell (2000). Influence of the 1997-98 El Niño on the surface chlorophyll in the California Current. Geophys. Res. Letters. 27: 2937-2940.