A Biooptical Model of Irradiance Distribution and Photosynthesis in Seagrass Canopies

Although extremely vulnerable to coastal eutrophication, seagrasses represent important structuring elements and sources of primary production in shallow waters. They also generate an optical signature that can be tracked remotely. Accurate knowledge of light absorption and scattering by submerged plant canopies permits the calculation of important plantand ecosystem-level properties, including rates of photosynthesis, vegetation abundance, and distribution. The objectives of this study were to develop a realistic, yet simply parameterized two-flow model of plane irradiance distribution through a seagrass canopy submerged in an optically active water column, to evaluate its performance against in situ measurements, and to explore the impacts of variations in canopy architecture on irradiance distribution and photosynthesis within the canopy. Allometric functions derived from leaf length–frequency data enabled simple parameterization of canopy architecture. Model predictions of downwelling spectral irradiance distributions in seagrass canopies growing in both oligotrophic and eutrophic waters were within 15% of field measurements. Thus, the model provides a robust tool for investigating photosynthetic performance of seagrass canopies as functions of water quality, depth distribution, canopy architecture, and leaf orientation. Model predictions of upwelling irradiance were less reliable, particularly in the upper half of the canopies. The model was more sensitive to leaf orientation than leaf optical properties, seabed reflectance, or the average cosine of downwelling irradiance. Better knowledge of leaf orientation appears to be a fruitful avenue for improving our understanding of the interaction between seagrasses and the submarine light environment. The distribution of radiant energy in plant canopies determines one of the fundamental interactions of biophysical ecology—that of energy exchange between photosynthetic organisms and their environment. Accurate knowledge of light absorption by plant canopies permits the calculation of important plantand ecosystem-level properties, including rates of primary production and, in the case of terrestrial plants, evapotranspiration (Nobel 1991; Campbell and Norman 1998). Knowledge of light scattering by plant canopies is crucial for remote sensing quantification of vegetation abundance and distribution, as well as for the development of inversion techniques to infer plant chemical composition important for ecosystem-scale estimates of plant growth and 1 Corresponding author ([email protected]). Acknowledgments Many thanks to all CoBOP participants for generous collaboration, assistance with fieldwork and warm camaraderie during five field campaigns to Lee Stocking Island. Special thanks to D. Kohrs, M. Cummings, S. Wittlinger, and S. Palacios for assistance with laboratory and field measures of seagrass density, leaf dimensions, and optical properties, and numerous discussions and suggestions that improved both the model and manuscript. E. Louchard and P. Reid provided carbonate sediment reflectance spectra; E. Boss and R. Zaneveld provided water column optical properties from Lee Stocking Island. J. Norris shared eelgrass morphometrics from Dumas Bay, Washington, collected under the auspices of the Washington State Department of Natural Resources. R. Maffione provided Hydrolight simulations used to parameterize water column variables. R. Maffione, C. Mobley, H. Dierssen, and L. Drake also provided important discussions and suggestions for improving both the model and the manuscript. A. Morel and two anonymous reviewers provided extensive, thoughtful comments that greatly improved the final product. Financial support was provided by the Environmental Optics program, Office of Naval Research, as part of the CoBOP DRI (award No. N00014-97-1-0032). biogeochemical flux (Jacquemoud et al. 1996; LaCapra et al. 1996; Broge and Leblanc 2000). Seagrass meadows represent an important structuring element and major source of primary production in shallow waters worldwide (Hemminga and Duarte 2000). They also provide a strong optical signature that can be tracked using satellite and airborne remote sensing (Armstrong 1993; Mumby et al. 1997; Chauvaud et al. 2001; Dierssen et al. 2003). Despite the high productivity of seagrass meadows, these marine angiosperms are vulnerable to the light-limiting effects of sediment loading and coastal eutrophication (Orth 1977; Cambridge and McComb 1984; Zimmerman et al. 1991; Tomasko and Lapointe 1994; Zimmerman et al. 1995). Limited understanding of many processes crucial to seagrass production often necessitates the development of semiempirical relationships with limited applicability to other species or locations. Consequently, models of light-limited seagrass productivity have been constructed using a variety of assumptions and different levels of mechanistic detail (e.g., Short 1980; Wetzel and Neckles 1986; Burd and Dunton 2001). Self-shading, which ultimately regulates the density of any plant canopy, has been incorporated into seagrass production models through a simple light attenuation coefficient derived by correlation from canopy height and shoot density (Short 1980). It has also been incorporated as an explicit feedback term affecting photosynthesis but without a direct link to the submarine light environment (Burd and Dunton 2001). Both approaches can provide useful predictions of seagrass productivity when there are sufficient data for accurate least-squares parameterization of the transfer coefficients for a particular population or environment. These models, however, are not easily generalized to other populations or situations without recalibration, which requires extensive collection of new data. Further, the formulations do 569 Irradiance in seagrass canopies not separate the effects of water column attenuation from those of the canopy. This hinders their utility for developing general water-quality criteria to manage submarine light environments with respect to submerged aquatic vegetation in heterogeneous and temporally variable coastal waters. Finally, they do not address the canopy-leaving irradiance, which is necessary for remote sensing of seagrass meadows. Radiative transfer theory provides a robust framework to develop more mechanistic models of the submarine light environment, system-level productivity, and remotely sensed reflectance of seagrass meadows in optically shallow waters. Exact solutions to the radiance transfer equations have been developed for natural waters in which the optical medium is a continuous material composed of randomly arranged scattering elements separated by large distances relative to the wavelength of light (Mobley 1994). Plant leaves, however, represent a dense packaging of optically active material, which violates the single-scattering assumptions of these exact solutions. Consequently, models of irradiance distribution in plant canopies must rely on more empirical relationships between leaf optical properties and light attenuation by the bulk canopy (Goudriaan 1988; Shultis and Myneni 1988; Ganapol and Myneni 1992). Although less elegant mathematically, the simulation of irradiance transfer by the use of two-flow equations provides a simple, quasimechanistic framework for understanding the relationship between water column optical properties, submerged plant canopies, and irradiance distribution in shallow waters. Ackleson and Klemas (1986) developed a two-flow model to calculate the submarine light field within a twolayered system composed of a homogeneous water column bounded below by a vertically homogeneous seagrass canopy. Zimmerman and Mobley (1997) extended that approach with a vertically layered model of downwelling irradiance distribution in seagrass canopies but ignored water column effects within the canopy and modeled the leaves as scalar irradiance collectors. The goals of this study were (i) to develop a more realistic, yet simply parameterized model of plane irradiance distribution through a vertically defined leaf canopy submerged in an optically active water column, (ii) to evaluate its performance against in situ measurements, and (iii) to explore the impacts of variations in canopy architecture on irradiance distribution and photosynthesis. Mechanistic characterization of the details of light absorption by vertically differentiated plant canopies helps to understand the environmental regulation of seagrass productivity. Understanding the factors that determine the emergence of upwelling irradiance from submerged plant canopies will be instrumental in creating remote sensing algorithms necessary to develop global inventories of submerged aquatic vegetation and protect these critical coastal resources (Dierssen et al. 2003). Material and methods The model—The model consisted of three modules simulating (i) vertical canopy architecture and leaf geometry, (ii) irradiance distribution within the simulated seagrass canopy, and (iii) canopy photosynthesis resulting from light absorption by the leaves. The optical medium (canopy plus water column) was divided into a vertical series of planeparallel, horizontally homogeneous slabs of finite thickness (Dz). Optical properties of each layer were based on canopy architecture, leaf optical properties, and the optical properties of the water column. Depth incremented positively downward, so that z 5 0 represented the top layer of the submerged seagrass canopy. The parenthetical notation (l, z) denotes wavelength and/or depth dependence of the specified terms. Symbol definitions and dimensions are provided in Table 1. Module i: Vertical canopy architecture and leaf geometry—A mathematical description of canopy architecture and leaf orientation was derived from morphometric analysis of 35 seagrass populations and some geometric reasoning. Leaf size–frequency distribution and shoot-specific leaf area were measured using shoots collected from 32 turtlegrass (Thalassia testudinum Banks ex König) populations near Lee Stocking Island, Bahamas (LSI) and three eelgrass (Zostera marina L.) populations from California and Washington, USA (Table 2). Shoot density was determined at each site by direct counts of all shoots within twenty randomly located 0.1-m2 quadrats. One shoot was collected from each quadrat for detailed measurement of shoot morphology and leaf optical properties. Lengths of all leaves on each shoot were measured to the nearest millimeter with a plastic meter tape. Leaf widths were measured to the nearest 0.01 mm using a digital caliper. Shoot-specific leaf area (Ls) was calculated as the sum of the one-sided area (length 3 width) of all leaves on each shoot. The vertical distribution of leaf biomass within each canopy was determined from the respective leaf length–frequency distributions assuming that leaves were vertically oriented and originated at the seafloor. Leaf area index (L) for the entire canopy was defined as the product of the shoot density and the one-sided leaf area per shoot (Ls). It was distributed through the canopy as a function of the relative amount of biomass [B(z)], such that l(z) represented the leaf area index at depth z: