Powell et al_2005 - 1568 Ecosystem and understory water and...

Info iconThis preview shows page 1. Sign up to view the full content.

View Full Document Right Arrow Icon
This is the end of the preview. Sign up to access the rest of the document.

Unformatted text preview: 1568 Ecosystem and understory water and energy exchange for a mature, naturally regenerated pine flatwoods forest in north Florida1 Thomas L. Powell, Gregory Starr, Kenneth L. Clark, Timothy A. Martin, and Henry L. Gholz Abstract: Eddy covariance was used to measure energy fluxes from July 2000 – June 2002 above the tree canopy and above the understory in a mature, naturally regenerated slash pine (Pinus elliottii Engelm. var. elliottii) – longleaf pine (Pinus palustris Mill.) flatwoods forest. Understory latent energy (λE ) and sensible heat (H) fluxes accounted for 45% and 55% of whole-ecosystem fluxes, respectively, with strong seasonal variation in the proportion of λE attributable to the understory. The partitioning of net radiation (Rnet) to λE and H also changed seasonally, with half-hourly mean ecosystem H in the winter peaking at 175 W·m–2, almost twice as large as λE . In contrast, half-hourly ecosystem λE and H remained almost equal throughout the day in July and August, with mean midday peaks of approximately 200 W·m–2. Maximum hourly evapotranspiration (ET) in the months of July and August was 0.32 and 0.29 mm·h–1 for 2000 and 2001, respectively. For a variety of environmental conditions, mean daily ET was approximately 2.7 mm in the summer and 1.3 mm in the winter. Annual ET for the first year was 832 mm, or 87% of annual precipitation (956 mm). Although leaf area index was higher in the second year, annual ET was only 676 mm, which is considerably lower than that of the previous year, but it still accounted for approximately the same proportion (84%) of the much lower annual precipitation (811 mm). Canopy conductance declined as soils dried, changing patterns of partitioning of Rnet to λE . Résumé : La méthode de covariance des turbulences a été utilisée pour mesurer les flux d’énergie de juillet 2000 à juin 2002 au-dessus du couvert forestier et au-dessus de la végétation du sous-bois d’une forêt mature d’origine naturelle, établie sur terrain plat et composée de pin d’Elliott typique (Pinus elliottii Engelm. var. elliottii) et de pin des marais (Pinus palustris Mill.). Les flux de chaleur latente (λE ) et sensible (H) de la végétation du sous-bois représentaient respectivement 45 et 55 % des flux de tout l’écosystème et de fortes variations saisonnières dans la proportion de λE ont été causées par la végétation du sous-bois. La répartition du rayonnement net (Rnet) entre λE et H a aussi varié selon les saisons avec des valeurs mi-horaires moyennes de H de l’écosystème atteignant un maximum de 175 W·m–2 pendant l’hiver, soit presque le double de λE . Au contraire, les valeurs mi-horaires de λE et H de l’écosystème sont demeurées presque égales tout au long des journées en juillet et août avec des valeurs maximales de mi-journée d’environ 200 W·m–2. L’évapotranspiration (ET) horaire maximale au cours des mois de juillet et août a atteint respectivement 0,32 et 0,29 mm·h–1 pour les années 2000 et 2001. Pour toute une gamme de conditions environnementales, la moyenne journalière de ET était d’environ 2,7 mm pendant l’été et 1,3 mm pendant l’hiver. La valeur annuelle de ET pour la première année était de 832 mm, soit 87 % des précipitations annuelles (956 mm). Même si la valeur de l’indice de surface foliaire était supérieure pendant la deuxième année, la valeur annuelle de ET a atteint seulement 676 mm, ce qui est considérablement plus faible que la valeur de l’année précédente. Cette valeur correspond environ à la même proportion (84 %) des précipitations annuelles (811 mm) de 2001 qui ont été beaucoup plus faibles qu’en 2000. La conductance du couvert forestier a diminué avec l’assèchement du sol, ce qui a changé le patron de répartition de Rnet sous forme de λE . [Traduit par la Rédaction] Powell et al. 1580 Received 20 June 2004. Accepted 18 March 2005. Published on the NRC Research Press Web site at http://cjfr.nrc.ca on 10 August 2005. T.L. Powell,2 G. Starr, K.L. Clark,3 T.A. Martin,4 and H.L. Gholz.5 School of Forest Resources and Conservation, University of Florida, P.O. Box 110410, Gainesville, FL 32611, USA. 1 Journal Series R-10682 of the Florida Agricultural Experiment Station, Gainesville, Florida. Present address: Smithsonian Environmental Research Center, Mail Code: DYN-3, Kennedy Space Center, FL 32899, USA. 3 Present address: USDA Forest Service, Northeastern Research Station, 11 Campus Boulevard, Ste. 200, Newtown Square, PA 19073, USA. 4 Corresponding author (e-mail: tamartin@ufl.edu). 5 Present address: Division of Environmental Biology, National Science Foundation, 4201 Wilson Boulevard, Arlington, VA 22230, USA. 2 Can. J. For. Res. 35: 1568–1580 (2005) doi: 10.1139/X05-075 © 2005 NRC Canada Powell et al. Introduction Pine flatwoods are a widespread ecosystem type on the southeastern US Coastal Plain, occupying approximately 5.35 × 106 ha (Smith et al. 2001). Flatwoods are particularly important in Florida, where they account for about half of the terrestrial landscape, in a mixture of about two-thirds pine uplands and one-third forested wetlands (Myers and Ewel 1990). Prior to European settlement, these ecosystems were characterized by open-canopy forests of mixed longleaf pine (Pinus palustris Mill.) and slash pine (Pinus elliottii Engelm. var. elliottii), with low, but dense shrub understories (Abrahamson and Hartnett 1990). Over the last century, the structure of the pine uplands has changed considerably, and much of the naturally occurring pine forests have been replaced by artificially regenerated plantations of slash pine, and to a lesser extent loblolly pine (Pinus taeda L.). These plantations are managed with varying degrees of silvicultural intensity (Brown 1996). Plantations differ in structure and function from their naturally regenerated predecessors in a number of ways. Compared with most naturally regenerated forests, pine plantations have higher stem density, higher tree canopy leaf area index (LAI), lower understory LAI, smaller mean tree size, and a smaller range of tree sizes and tree ages. In addition, the low-intensity, high-frequency fires that characterized presettlement forests are essentially excluded from pine plantations. The effects of these contrasting structural and functional attributes on the energy balance and water fluxes of the landscape are not known, although many public and some nonindustrial private forest landowners in the region are increasingly turning to natural regeneration and uneven-aged silviculture on their lands (Owen 2002). Variations in water flux and energy balance across space and time are a function of interactions among environmental conditions, ecosystem characteristics, and land management regimes. A complex mosaic of land uses within landscapes induces energy gradients that influence local and regional water dynamics (Baldocchi et al. 1996). Evaluating the effects of even-aged forest management on ecosystem and regional water budgets on the southeastern Coastal Plain has been the focal point of considerable research (Riekerk 1982, 1989; Golkin and Ewel 1984; Liu 1996; Liu et al. 1998; Martin 2000; Gholz and Clark 2002). This research has shown that evapotranspiration (ET) is the largest component of the flatwoods water balance and accounts for 60%–90% of annual precipitation. Furthermore, ET plays a large role in surface and ground water dynamics because shifts in the ratio between ET and precipitation can change the amount of precipitation that ends up in groundwater storage or streamflow (Allen 1982). The goal of this study was to quantify ecosystem water and energy budgets for a mature, naturally regenerated pine flatwoods ecosystem in north-central Florida, and to understand the key environmental variables that determine these budgets. The research had three primary objectives: (1) to quantify the water budget for this forest over a 2-year period; (2) to determine the dominant environmental factors controlling the seasonal partitioning of net radiation (Rnet) into the main energy budget components of latent energy flux (λE) and sensible heat flux (H); and (3) to determine the role of understory vegetation in ecosystem exchanges of water and energy. 1569 Table 1. Characteristics of the naturally regenerated Pinus palustris – Pinus elliottii stand in the Austin Cary Memorial Forest. Stand area Stem density Canopy height Tree DBH (2001) Tree basal area (2001) Tree canopy LAIa Understory LAIa 41 ha (95% pine uplands, 5% cypress wetlands) 325±36 ha–1 22.1±0.55 m 25.7±1.17 cm 18.0±0.95 m2·ha–1 2000–2001: summer, 3.4; winter, 2.6 2001–2002: summer, 3.8; winter, 2.6 Jan. 2001: Serenoa, 0.7; Ilex, 0.2 Feb. 2002: Serenoa, 1.0; Ilex, 0.2 Note: Values are mean ± 1 standard error. a All-sided leaf area index. Study site and methods This study was conducted on the University of Florida’s Austin Cary Memorial Forest (ACMF), 15 km northeast of Gainesville, Alachua County, Florida, USA (29°44′N, 82°09′30′ ′W). The elevation of the site is 50 m and the topography is flat. The study stand (Table 1) was a naturally regenerated, 41-ha, mixed slash pine and longleaf pine stand (64% and 36% of the stand basal area, respectively). Prior to state purchase in 1936, the forest had been selectively harvested for timber and otherwise utilized as grazing land. A current management objective is ultimately to restore elements of forest structure and function typical prior to European settlement, including large, widely spaced trees; a vigorous understory plant community; and frequent, low-intensity fires. The stand was thinned in 1991, removing 27% of the basal area, and the most recent prescribed fire occurred 4 years prior to this study. The stand is well stratified, with a 22.1 m tall, relatively open tree canopy 15 m above a 1.5 m tall understory. Tree ages ranged from 20 to 80 years in 2001, with a mean age of 60 years. The understory was well developed, consisting of naturally occurring native species dominated by saw palmetto (Serenoa repens (Bartr.) Small), gallberry (Ilex glabra (L.) Gray), wax myrtle (Myrica cerifera L.), and wiregrass (Aristida stricta Michx). The soils of this site are poorly drained Ultic Alaquods that have a discontinuous spodic horizon (30–60 cm depth) and a deeper argillic horizon (100–140 cm depth, Gaston et al. 1990). During the study period, the watertable ranged from 1.3 m depth in September 2000 to nearly 3 m in depth in July 2000 and June 2001 (Fig. 2E), with the average considerably deeper than the long-term average of 1–2 m (Abrahamson and Hartnett 1990). Long-term (1961–2004) mean annual precipitation recorded 5 km from the site was 1259 mm, with 52% falling from June to September (National Climatic Data Center 2004). Annual precipitation during the study period was 956 and 811 mm during the first and second years, respectively (Fig. 1A). Long-term mean maximum and minimum temperatures for the months of January and July were 19.0 °C and 6.1 °C, and 32.6 °C and 21.8 °C, respectively (National Climatic Data Center 2004). During the study, mean maximum and minimum temperatures for January and July were 17.6 °C and 5.6 °C, and 31.8 °C and 22.9 °C, respectively (Fig. 1B). © 2005 NRC Canada 1570 Can. J. For. Res. Vol. 35, 2005 120 A 100 80 60 40 20 0 max. o Mean weekly Tair ( C) Weekly precipitation (mm) Fig. 1. (A) Weekly precipitation, (B) maximum and minimum air temperature, and (C) volumetric soil water content at 10 cm depth for the two measurement years. 30 20 10 min. B 0 Soil water content (at 10 cm, v/v) 0.25 C 0.20 0.15 0.10 0.05 0.00 07/00 10/00 01/01 04/01 Energy flux measurements Net fluxes of latent heat (λE) and sensible heat (H) were measured using standard eddy-covariance methods (e.g., Clark et al. 1999, 2004). λE was estimated as [1] λE = ρa w ′ c′ where ρa is the density of air, w′ is the instantaneous deviation of vertical wind speed from the mean vertical wind speed (w′ = w − w ), c′ is the instantaneous deviation of water vapor concentration from the mean water vapor concentration (c′ = c − c), and the overbar denotes a time averaged, 200-s running mean. A 30-m walk-up scaffolding tower was erected in the center of the stand, extending 8 m above mean canopy height. A mobile 3-m antenna tower was also erected to measure aboveunderstory fluxes (λEu and Hu). We assumed that λEu and Hu fluxes transferred across a horizontal plane existing between the bottom of the canopy and the understory (Baldocchi and Vogel 1996, 1997). The understory tower was moved be- 07/01 10/01 01/02 04/02 07/02 tween three fixed locations (30–40 m from the tower) to dampen local effects of understory heterogeneity, and data were pooled for analysis. Wind speed was measured with 3dimensional sonic anemometers mounted 2 m above the top of the canopy tower and 1 m above the understory tower (Windmaster Pro, Gill Instruments Ltd., Lymington, UK). Closed-path infrared gas analyzers (IRGA) (LI-6262, LICOR, Inc., Lincoln, Nebraska) were used to measure H2O concentrations at 10 Hz. The gas samples were pumped at 6.0 L·min–1 from the anemometers to the IRGAs through 4 mm i.d. Teflon® coated tubes. Mean gas sample lag times between the anemometers and the IRGAs were approximately 10 and 12 s for the canopy and understory towers, respectively. Nitrogen was run at a rate of 0.10 L·min–1 through a desiccant and into the IRGA reference cells to establish a water-free baseline. The IRGAs were calibrated every 2 to 3 days using a dew point generator (LI-610, LI-COR, Inc.). Flux calculation software was used to rotate the horizontal wind velocities to obtain turbulence statistics perpendicular © 2005 NRC Canada Powell et al. to the local streamline using Reynolds detrending (200-s constant, Katul et al. 1999). Net scalar fluxes were averaged at half-hour intervals. Fluxes were collectively corrected for attenuation of the gas concentrations in the sampling tube, the nonideal frequency response of the LI-6262, and sensor separation loss using transfer functions (Moncrieff et al. 1997). The data were screened to eliminate spurious or incomplete half-hourly data resulting from system malfunction or environmental disturbance. Half-hour values were eliminated if they met any of the following criteria: (1) 10-Hz data were incomplete because of interruptions for calibration or maintenance, (2) data were collected during rain events, (3) data were collected when atmospheric conditions were stable, leading to poor coupling between canopy air and atmospheric air (we defined stable conditions as half hours with friction velocity (u*) <0.2 m·s–1 (Goulden et al. 1996; Clark et al. 1999, 2004)), or (4) 10-Hz data exhibited unusually high or unusually low variation. In total, ecosystem fluxes were measured ~70% of the time and about one-third of those data were eliminated by the screening criteria. There was no apparent temporal bias in the distribution of screened data. Water and energy budget calculations Interception (I) was calculated by subtracting throughfall (TF) and stemflow (SF) from precipitation and was measured on an event basis periodically in the summer and winter to include seasonal variation in LAI. TF was collected in twenty 1 m long (808 cm2) troughs that were slightly angled and drained by funnels into 4-L containers. TF depth was determined by dividing the volume of water collected in each trough by the collection area. SF was estimated as 1% of precipitation (P) when the event was >2.5 mm (Allen and Gholz 1996), otherwise SF was assumed to be negligible. Interception for rainfall events in which TF was not measured was estimated from a regression of (TF + SF) on P. Annual ET was estimated by summing half-hourly abovecanopy eddy covariance ecosystem latent heat flux (λEe ) values. For half hours with missing data, λEe was estimated by substituting Rnet, vapor pressure deficit (VPD), aerodynamic conductance (ga), and canopy conductance (gc) into the Penman–Monteith equation (Monteith 1965). Rnet and VPD were obtained from the weather station, and ga was calculated as ga = u*2/uz (Kelliher 1993; Baldocchi and Vogel 1997), where uz is horizontal wind speed at the sonic anemometer height (30 m). If sonic anemometer data were not available for a time period, the median observed ga value was used. A model was used to simulate gc from measured Rnet and VPD data (Jarvis 1976; Martin et al. 1997): [2] $ gc = f (Rnet ) × f (VPD) $ where gc is the predicted bulk canopy conductance, and ƒ(Rnet) is a function relating gmax to Rnet, where gmax is defined as the hypothetical gc under a given Rnet when there is no reduction in gc due to VPD. A vapor pressure deficit function, f(VPD), derived from boundary line analysis (Martin et al. 1997) was then used to reduce gmax and produce the final estimate of gc. We parameterized the model with a data set of measured bulk gc calculated by substituting meteorological data, ga, and λEe into the inverted Penman–Monteith equation (Baldocchi and Vogel 1997; Martin et al. 1997). To 1571 keep the gc model simple and parsimonious, we designed the model to only capture typical variation in gc, and therefore eliminated unusually high or low gc values from the model parameterization data set using the following criteria: (1) parameterization data were restricted to those half hours when λEe was within 50% of λEe predicted from the regression between λEe and Rnet (40% of parameterization data); (2) gc values >0.012 m·s–1 were eliminated (<4% of parameterization data); (3) data were screened when Rnet < 200 W·m–2 and gmax > 0.002 m·s–1 (<1% of parameterization data). When the gap-filling model was compared with a validation eddy-covariance data set (n = 829 randomly selected half hours) that was not used for model parameterization, residuals showed slight underprediction at very high VPD (>2.5 kPa) and very low VPD (<0.5 kPa), but the measured and modeled λE sums were within 10% of each other. When the canopy was wet, the Penman–Monteith equation was reduced to the Penman equation (Penman 1948) by removing gc from the equation, assuming water only evaporated from exterior surfaces until the canopy dried. For gap-filling λEe during rain events, the Penman model was used for half hours that had precipitation and subsequent half hours until the Penman model evaporated the depth of I. Once I was evaporated, the model incorporating the Penman–Monteith equation and eq. 2 resumed. To facilitate analysis of seasonal variation in λEe and ecosystem sensible heat flux (He), we created standardized monthly values of these variables. We performed monthly regressions of λEe and He versus Rnet, and we used the predictions at Rnet = 500 W·m–2 as the standardized values. To estimate the relative importance of gc and ga for controlling evapotranspiration, the decoupling coefficient Ω was calculated as described by Jarvis and McNaughton (1986). Biomass and leaf area estimates A 50 m × 50 m inventory plot was randomly located in each of four quadrants around the tower. The inventory plots were used for measurements of tree diameter and height, canopy and understory LAI, and throughfall. LAI was calculated from annual litterfall collected monthly from ten 1-m2 litter traps randomly located within the inventory plots following the method described in Martin and Jokela (2004). Litterfall was dried for 3 days at 70 °C in an oven, and needles were separated and weighed to the nearest 0.1 g. The needlefall year was defined as March 1 through February 28, since new needle flushing begins in March. A needlefall curve was integrated with a needle-accretion curve calculated from a logistic equation (Gholz et al. 1991; Dougherty et al. 1995) to estimate seasonal changes in canopy foliage biomass. Multiplication by the appropriate specific leaf areas of each cohort produced area equivalents (Dalla-Tea and Jokela 1991; Gholz et al. 1991). Allometric equations were used to determine all-sided LAI of understory saw palmetto from measurements of frond blade length and rachis length made in a 25 m × 25 m subplot at the center of each inventory plot (Gholz et al. 1999). Gallberry LAI was estimated using measurements of total stem height (centimetres) of individual plants made in twenty 1-m2 subplots distributed randomly in census plots, using a previously derived equation based on destructive sampling: © 2005 NRC Canada 1572 Can. J. For. Res. Vol. 35, 2005 Fig. 2. Monthly values of ecosystem latent energy (λ Ee) (A) and sensible heat (He) (B) fluxes standardized at a net radiation (Rnet) of 500 W·m–2, derived from monthly regressions of λ Ee and He against Rnet; mean monthly midday (1000–1400 hours) values of the Ω decoupling coefficient (C), monthly needlefall (D), and watertable depth (E). leaf area (cm2/plant) = (total height × 6.652) – 110.2 (n = 30, H.L. Gholz, unpublished data). Meteorological and soil measurements Meteorological data were continuously collected during the study period by scanning standard meteorological sensors every 5 min and then averaging every half hour with a datalogger (Easylogger EL824-GP, Omnidata International, Ogden, Utah). Sensors were mounted on top of the tower for measuring incident solar radiation, photosynthetically active photon flux density (LI-200 and LI-190, respectively, LICOR, Inc.), net radiation (Q7, Radiation and Energy Balance Systems, Inc., Seattle, Washington), wind speed and direction (3001-5, R. M. Young Company, Traverse City, Michigan), temperature and relative humidity (HMP 45C, Vaisala, Inc., Helsinki, Finland), and precipitation (tipping bucket, Sierra Misco, Inc., Berkeley, California). Watertable depth was measured 15 m from the instrument tower with a Steven’s water depth gage (F-68, Leupold and Stevens, Inc., Beaverton, Oregon). Volumetric soil moisture content, θ, and soil temperature, Ts, were monitored with combination thermistor – dielectric capacitance probes (Hydra, Stevens Vitel, Inc., Chantilly, Virginia). Ecosystem soil heat flux (G) was estimated by averaging soil heat flux measurements calculated from three soil heat flux plates (HFT-3.1, Radiation and Energy Balance Systems, Inc.) buried 10 cm below the soil surface in three separate locations, 8 m from the tower. The soil heat flux for each sensor was calculated as the sum of soil heat flux measured at sensor depth (10 cm) and the energy stored in the soil above the sensor, S (G = G10cm + S). Soil bulk density of 0.72 g·cm–3 (Schmalzer et al. 2001) and a soil heat capacity of 840 J·kg–1·K–1 were used with local half-hourly measurements of Ts and θ to calculate S (HFT3.1 technical reference). Statistical analysis Regression analysis using SigmaPlot 5.0 Regression Wizard (SPSS Science, Chicago, Illinois) and the GLM procedure of © 2005 NRC Canada Powell et al. 1573 Table 2. Energy budget equations and above and below canopy flux equations for net radiation (Rnet), latent energy (λ E), sensible heat (H), and soil heat flux (G). Equation Above-canopy energy balance, October 2000 – May 2002 Above-canopy latent energy flux 2000–2001 September 2000 January 2001 2001–2002 September 2001 January 2002 Above-canopy sensible heat flux, 2000–2002 Understory fluxes 2000–2001a 2001–2002a R2 p value n (λ Ee + He + G) = 0.80Rnet – 17.0 0.83 <0.0001 6962 λ Ee = 0.29Rnet + 17.5 λ Ee = 0.42Rnet + 13.7 λ Ee = 0.18Rnet + 12.2 λ Ee = 0.29Rnet + 6.1 λ Ee = 0.34Rnet + 14.1 λ Ee = 0.26Rnet + 5.7 He = 0.45Rnet – 19.0 0.51 0.65 0.42 0.59 0.64 0.58 0.81 <0.0001 <0.0001 <0.0001 <0.0001 <0.0001 <0.0001 <0.0001 4021 278 340 4666 374 311 8757 λ Eu = 0.51λ Ee Hu = 0.55He – λ Eu = 0.43λ Ee Hu = 0.45He – 0.62 0.81 0.52 0.77 <0.0001 <0.0001 <0.0001 <0.0001 2102 2102 1553 1582 + 14.2 7.2 + 19.4 3.9 Note: All units are watts per square metre. September and January are shown to represent the highest and lowest months associated with the energy components of the system. a Pooled across three understory tower locations. SAS 8.1 (SAS Inc., Cary, North Carolina) was conducted to explore relationships between λE and H and environmental variables. Analysis of covariance with indicator variables was used to detect differences between seasonal regressions. Census plot plant measurement data were analyzed using Statistica (StatsSoft, Inc., Tulsa, Oklahoma) with the nonparametric Kruskal–Wallis test to determine whether there were significant interactions (p = 0.05) between plots and individual samples within the plots (Stevens 1996). No significant plot effects were found, so all plot data were pooled. Results Meteorological data During the study, Florida was experiencing a “100-year drought”, with total precipitation of 956 and 811 mm during the first and second years, respectively, with approximately 60% falling during July to October each year (Fig. 1A). There were 86 days when P > 1 mm in the first year and 78 days when P > 1 mm during the second year. Volumetric soil water content (θ) at 10 cm depth varied widely, from field capacity (25% for the sandy soils at this site) to as low as 3.5% after long periods with low precipitation (Fig. 1C). The low θ values of 3.5% are equivalent to soil matric potential of approximately –0.78 MPa, based on soil moisture release curves generated for local Spodosols (H.L. Gholz unpublished data). Needlefall and leaf area dynamics In addition to the normal fall pulse of needlefall (Gholz et al. 1991), there were premature needlefall pulses in May and June of 2000 and 2001, where 50% of the annual needlefall occurred (Fig. 2D). Canopy LAI (all-sided) had a seasonal trend with peak values between 3.4 and 4.0 in August and minimum values between 2.2 and 2.6 in March. Understory LAI was 0.9 in January 2001 and 1.2 in January 2002 (Table 1). Exchanges of momentum and energy The relationship between friction velocity (u*, metres per second) and horizontal wind speed (uz, metres per second) was linear (u* = 0.241uz – 0.086; R2 = 0.72, n = 11997). This relationship included data from all wind directions and indicates that eddies penetrated the open canopy in a relatively uniform manner. Net radiation was linearly related to incident solar radiation, Rg (Rnet = 0.77Rg – 14.2; R2 = 0.98, n = 4373). A mean albedo of 0.23 was calculated from the complement of the slope of this relationship. A comparison between the sum of λ Ee, He, and G and Rnet revealed that the annual energy balance at this site closed to within 20% (Table 2, Fig. 3). Ecosystem λE was linearly related to Rnet and accounted for approximately 30% of Rnet (Table 2, Fig. 4A). The relationship between λEe and Rnet was significantly different between the 2 years (slope p = 0.04, intercept p < 0.0001). Ecosystem sensible heat flux (He) was also linearly related to Rnet, and, on average, accounted for about 40% of Rnet. This relationship did not vary between the two measurement years (Table 2, Fig. 4B). Bimodal patterns were observed each year in time series of λEe and He standardized to Rnet = 500 W·m–2 (Figs. 2A and 2B). A minor peak in standardized λEe occurred in April, and a major peak occurred in September of each year. A minor trough occurred in June and July 2001, and a major trough occurred each January. The September peaks coincided with maximum LAI in each year, while each trough was associated with a major pulse of litterfall (Fig. 2D). Trends in standardized He were generally complementary to variation in standardized λEe (Fig. 2B). The partitioning of Rnet to λEe varied with both VPD and soil moisture (Fig. 5). As VPD increased from 0.00 to 0.75 kPa, the proportion of R net dissipated as λ E e increased linearly from about 17% to 30%. Under dry soil conditions (θ at 10 cm depth <6%, soil water potential < –0.45 MPa), λ Ee/Rnet remained steady at 0.30 until VPD reached © 2005 NRC Canada 1574 Can. J. For. Res. Vol. 35, 2005 Fig. 3. Annual energy balance the for the University of Florida’s Austin Cary Memorial Forest during October 2000 – May 2002. For clarity, a 15% random sample of the data points is plotted. See Table 2 for corresponding equations and statistics. 1:1 900 700 900 600 A 500 700 400 300 200 100 1:1 2000 - 2001 2001 - 2002 800 λEe (W·m–2) –2 λE + H + G (W·m ) 800 Fig. 4. (A) The relationship between net radiation (Rnet) and ecosystem latent energy flux (λ Ee) for 2000–2001 (yr1) and 2001– 2002 (yr2). The slopes and intercepts were significantly different between years. (B) The relationship between Rnet and ecosystem sensible heat (He) for all data pooled over both years. There was no significant difference in this relationship between years. For clarity, a 10% random sample of the data points is plotted. See Table 2 for corresponding equations. 600 500 400 0 300 yr1 200 0 yr2 100 100 200 300 400 500 600 700 800 900 –2 0 R net (W·m ) 0 Conductances and ET modeling We used a screened subset of the eddy-covariance data to investigate “typical” variation in canopy conductance (gc) with environmental drivers, and to create a model for filling 100 200 300 400 500 600 700 800 900 –2 Rnet (W·m ) 900 B 1:1 800 700 He (W·m–2) 2.0 kPa, at which point the ratio began declining. Under wet soil conditions, λ Ee/Rnet increased linearly to 0.40 at a VPD of 1.25 kPa and then remained relatively constant. Weekly non-light-limited gc (photosynthetically active photon flux density >1500 µmol·m–2·s–1) was positively but weakly correlated with θ (gc = 0.003 + 0.0316θ; R2 = 0.21). Diurnal patterns of ecosystem and understory λE and H in winter months (December and January) and summer months (July and August) highlighted seasonal shifts in energy partitioning (Fig. 6). In December and January, He dominated ecosystem energy balance, with mean midday values of 175 W·m–2, almost twice as large as λEe (Fig. 6A). In contrast, mean λ Ee and He remained almost equal throughout the day in July and August (Fig. 6B). Understory energy partitioning was much less variable over seasons, with λEu and Hu remaining approximately equal throughout the day in both summer and winter months (Fig. 6). Understory (understory vegetation plus litter layer and soil) latent energy (λEu) and sensible heat (Hu) fluxes were a significant proportion of λEe and He (Figs. 6 and 7). On an annual basis, λEu and Hu accounted for 45%–55% of wholeecosystem fluxes (Table 2). Maximum half-hourly λEu and Hu were approximately 200 and 300 W·m–2, respectively. There was a major contrast in the proportion of λEe accounted for by λEu in midsummer (July and August) versus midwinter (December and January). In midwinter λEu dominated λEe , with understory latent heat loss accounting for 75% to almost 100% of ecosystem water loss at some points during the diurnal cycle (Fig. 6A). During midsummer, λEu accounted for about 50% of λEe throughout most of the day (Fig. 6B). The proportion of He accounted for by Hu remained steady throughout the year (Fig. 6). 600 500 400 300 200 100 0 0 100 200 300 400 500 600 700 800 900 –2 Rnet (W·m ) in missing values from the eddy-covariance time series (Table 3). Aerodynamic conductance (ga) in this data set ranged from 0.02 to 0.33 m·s–1 and had a weak linear relationship with uz (Table 3). Canopy conductance (gc) was highly variable, but declined nonlinearly from 0.048 to 0.001 m·s–1 with increasing VPD (Table 3). Maximum canopy conductance, the value of gc that would be expected without VPD limitations, increased linearly with Rnet (Table 3). Evapotranspiration estimated by substituting the modeled gc values into the Penman–Monteith equation agreed well with eddy-covariance results for a validation data subset (R2 = 0.61, p < 0.0001, n = 1657). A sensitivity analysis revealed that annual ET estimates varied by less than 2% when the gap-filling model parameters (eq. 2) were individually changed by ±10%. © 2005 NRC Canada Powell et al. 1575 Fig. 5. Relationship between ecosystem latent energy flux (λ Ee) normalized by net radiation (Rnet) and vapor pressure deficit for dry soil conditions (volumetric soil water content (SWC) <6%) and moist soil condition (SWC >6%). 0.50 0.45 SWC < 6% SWC > 6% Fig. 7. Proportion of ecosystem latent energy (λ Ee) and sensible heat (He) accounted for by the understory latent energy (λ Eu) (A) and sensible heat (Hu) (B) for November 2000 – June 2001 (1, closed symbols) and July 2001 – May 2002 (2, open symbols). For clarity, a 20% random sample of the data points is plotted. See Table 2 for corresponding equations and statistics. 350 300 0.35 0.30 1:1 250 200 λEu (W·m–2) λE/Rnet 0.40 Nov. 2000 - June 2001 July 2001 - May 2002 0.25 0.20 A 1 2 150 100 0.15 50 0.10 0 0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 Vapor Pressure Deficit (kPa) -50 -50 Fig. 6. Diurnal course of ecosystem and understory energy flux components for (A) midwinter months (December and January) and (B) midsummer months (July and August). A Hu λEe 100 He Hu (W·m–2) Energy Flux (W·m–2) 150 Nov. 2000 - June 2001 July 2001 - May 2002 1:1 B 1 200 2 150 100 50 0 50 -50 0 -50 -50 0 50 100 150 200 250 300 350 400 450 500 He (W·m–2) 250 B 200 Energy Flux (W·m–2) λEe (W·m–2) 250 200 λEu 50 100 150 200 250 300 350 400 450 500 350 300 250 0 150 100 50 0 -50 0600 0800 1000 1200 1400 1600 1800 2000 2200 Time of Day The mean value of the decoupling coefficient Ω varied monthly throughout the study period, ranging between 0.07 and 0.22 (Fig. 2C), indicating strong coupling or stomatal control of transpiration. Variation in Ω followed fluctuations in the watertable depth (Fig. 2E). Monthly mean Ω was positively but weakly correlated with θ at 10 cm depth (Ω = 0.05 + 0.966θ; R2 = 0.23, data not shown). Precipitation, evapotranspiration, and canopy interception Total precipitation (P) was 956 and 811 mm during the first and second years of the study, respectively, with approximately 60% falling from July to October each year. Annual evapotranspiration (ET) was estimated to be 832 and 676 mm for the 2 years. Throughfall (TF) was measured during the first year and the sum of TF plus stemflow (SF) was strongly related to P (TF + SF = 0.82P – 0.1897; R2 = 0.99, p < 0.0001). This relationship can also be expressed in terms of canopy interception loss (I = 0.1761P + 0.1897; R2 = © 2005 NRC Canada 1576 Can. J. For. Res. Vol. 35, 2005 Table 3. Equations defining the relationship between aerodynamic conductance (ga) and horizontal wind speed (uz), and functions for modeling bulk canopy conductance (gc), where predicted $ gc = f ( VPD) × f ( Rnet ). Equation Aerodynamic conductance VPD functiona Rnet function R2 p value n ga = 0.033uz – 0.061 ƒ(VPD) = e(–0.7727VPD) ƒ(Rnet) = 3 × 10–5Rnet + 0.0021 0.22 — 0.5 <0.0001 — <0.0001 4599 301 2933 a VPD, vapor pressure deficit; Rnet, net radiation. This function was fitted through the top 10% of the data set to define the upper boundary rather than to describe the mean response, and therefore regression statistics are not reported. 0.86, p < 0.0001), where the intercept represents the canopy storage capacity in millimetres. Annual I for this stand was estimated to be 205 mm for year 1 and, by applying the same relationship, 177 mm for year 2. Over the course of the study, mean daily ET was approximately 2.7 and 1.3 mm·day–1 during summer (July–September) and winter (December–February), respectively. Mean hourly ET peaked at 0.32 mm·h–1 during the summer of 2000 and 0.29 mm·h–1 during the summer of 2001. The mean diurnal course of ET during the summer of the second year was asymmetrical, with a distinct midafternoon depression. The diurnal course of ET during the winter of both years was symmetrical, occurring over a 12-h period and peaking at 0.23 mm·h–1 at 1300 hours. Weekly ET tracked seasonal fluctuation in Rnet more closely than did LAI or environmental conditions. There did not appear to be a watertable threshold below which ET rates were markedly reduced, yet when λ E was normalized by Rnet, it did tend to decrease over time as the watertable continued to decline over the study period (Fig. 2). Discussion During the study, Florida was experiencing a “100-year drought”, with annual precipitation falling about 30% below long-term averages. Typically, pines in these ecosystems shed most of their needles in the mid to late fall, with a much smaller needlefall peak in May to June (e.g., Dalla-Tea and Jokela 1991; Gholz et al. 1991). However, up to half of the needlefall occurred in May–June during this study, with only about a quarter during the normal fall period. The premature needlefall caused summer LAI to be lower than usual and the seasonal range to be consequently dampened (about 30% compared with 40% in studies from wetter years, Gholz et al. 1991). Within-year variation in ET and energy exchange for other local ecosystems are closely linked to LAI (Liu 1998; Liu et al. 1998; Gholz and Clark 2002), which suggests that seasonal fluctuations over time in forests in the region may be high as they experience and then recover from periodic droughts. Over the study period, mean albedo was 0.23, which is at the low end of the range reported for other conifer forests (Jarvis et al. 1976; Baldocchi and Vogel 1996). This value is higher than the 0.18 reported for nearby closed-canopy plantations (Gholz and Clark 2002). In fact, it is closer to that of a nearby 2-year-old stand (Gholz and Clark 2002), probably because of the relatively low canopy LAI and greater radiation interception by understory vegetation (which is similar to that regrowing in the understory of the 2-year-old stand). The degree of energy budget closure is one way to assess the accuracy of component measurements in energy-balance studies. If measurements are taken without error, the sum of the energy dissipation terms (λ Ee, He, and G) should equal net energy input (Rnet), if it is assumed that energy storage in canopy air or biomass is negligible. On average, summed measurements of λ Ee, He, and G in our study were about 20% lower than Rnet (Table 2, Fig. 3). Similar departures from full energy budget closure were found in a naturally regenerated, semi-arid ponderosa pine (Pinus ponderosa) forest in Oregon (Anthoni et al. 1999), which also has an open canopy. While the 20% departure from closure is of similar magnitude to the potential error of measurement for Q7 Rnet sensors (±20%, Twine et al. 2000), it is also possible that the position of the Rnet sensor, 8 m above the mean canopy height, may not have been high enough to obtain a representative spatial characterization of Rnet for the entire flux footprint area. Additional error in our hourly flux estimates may have occurred because we did not measure heat storage in the canopy air or biomass, which can approach 100 W·m–2 under some midday conditions (Baldocchi and Vogel 1997). Comparisons of λ E and H between this ecosystem and other pine forests suggest similarities in the partitioning of net radiation across greatly contrasting environments. For example, a ponderosa pine forest in Oregon (Anthoni et al. 1999), a maritime pine (Pinus pinaster) forest in Portugal (Berbigier et al. 1996), a Scots pine (Pinus sylvestris) forest in Siberia (Kelliher et al. 1998), and a boreal jack pine (Pinus banksiana) forest in Saskatchewan (Baldocchi and Vogel 1996) all utilized 27%–36% of Rnet for λ Ee, which is similar to both the naturally regenerated stand in this study as well as the adjacent pine plantations (Table 4). The relatively low partitioning of available energy to λ Ee is typical of pine forests, which generally have low LAI and low leaflevel stomatal conductance, in contrast to many temperate broad-leaved forests, which tend to have higher LAI and leaf-level stomatal conductance. As a result, many temperate broadleaf forests partition 70% or more of Rnet to λ Ee (Baldocchi and Vogel 1996; Schmid et al. 2000; Wilson et al. 2002). Comparing the local Florida pine stands with contrasting management histories indicates that the ACMF stand and the plantations both utilized similar amounts of Rnet for λ Ee. At an Rnet of 500 W·m–2, λ Ee in the current study was 158 W·m–2 (averaged over the 2 years) compared with 164 W·m–2 in surrounding plantation stands (Gholz and Clark 2002). However, smaller seasonal differences in λ Ee were observed for the ACMF in comparison with the closed-canopy plantations. For example, summertime λ E (normalized to Rnet = 500 W·m–2) © 2005 NRC Canada Powell et al. 1577 Table 4. Water vapor exchange for present study and other selected sites. Forest type LAI (all-sided) Stems·ha–1 λ Ee (% of Rnet) λ Eu (% of λ Ee) Mean ET (mm·day–1) Reference Florida pine flatwoods Naturally regenerated, summer Naturally regenerated, winter Pinus elliottii plantation Rotation-aged, summer Rotation-aged, winter Pinus taeda plantation, wintera Pinus banksiana 3.4 2.6 325 325 33 28 50–60 70–85 2.7 1.3 This study This study 6.6 4.0 6.9 4.0 1300 1300 1150 1875 38 27 — 36 — — — 10–40 3.6 2.0 1.8 0.5–2.5 Pinus ponderosa 3.2 620 — Pinus pinaster Siberian Pinus sylvestris Temperate deciduous — 4.4 9.8 300 290 — 31 30 77 44 (early spring), 22–33 (summer) 15 54 5 Gholz and Clark 2002 Gholz and Clark 2002 Martin 2000 Baldocchi and Vogel 1996, 1997 Anthoni et al. 1999; Law et al. 2000 Berbigier et al. 1996 Kelliher et al. 1998 Baldocchi and Vogel 1996 ~1.6–1.7 2.1 — — Note: “—” indicates that data are not available. LAI, leaf area index, λEe, ecosystem latent energy flux; λEu, understory latent energy flux; Rnet, net radiation, ET, evapotranspiration. a Sapflow measurements on overstory only. in the plantation was 40% greater than wintertime λ E, while the summertime λ E for the ACMF was only about 25% greater than wintertime λ E. The smaller range for the ACMF was likely due to drought effects on gc, which would have been affected by reductions in both LAI and leaf-level stomatal conductance. Seasonal differences in standardized He for the plantations (11%) were similar to the seasonal range for the ACMF (12%–23%). The relatively few studies that have quantified understory energy balance have shown substantial variation in the proportion of ecosystem energy flux contributed by the understory (Table 4; see also review by Black and Kelliher 1989). As might be expected, forests with sparse or no understory tend to have a smaller contribution of λEu to λEe than forests with well-developed understory vegetation. For example, of the Douglas-fir (Pseudotsuga menziesii) stands reviewed by Black and Kelliher (1989), those with “sparse” or no understory vegetation had λEu λEe ratios ranging from 0.03 to 0.21, versus 0.30 to 0.65 for stands with understory projected LAI of 1.0 to 3.0. The stand in the present study would probably fall into the latter category, with understory all-sided LAI of 0.9 to 1.2, and λEu λEe of about 0.50. To some extent, the relationships among overstory LAI, understory LAI, and λEu λEe may be homeostatic. For instance, Roberts (1983) suggested that similarities in annual evapotranspiration among forest stands in Europe with widely varying overstory LAI may be partially attributable to compensatory water loss from understory vegetation in stands with lower overstory LAI. The λEu λEe ratio in the current study changed seasonally, with λEu contributing as much as 85% of ecosystem λE in the winter and 50%–60% in the summer (Fig. 6). Studies in other open-canopy pine forests have also shown seasonal variation in λEu λEe . The λEu λEe ratio in the ponderosa pine forest studied by Law et al. (2000) declined from 0.44 in spring to 0.22–0.33 in summer. They attributed this change to decreased soil evaporation from the sparsely vegetated understory as drought developed. In contrast, Black and Kelliher (1989) cited several studies in Pinus and Pseudotsuga for- ests that demonstrated increases in λEu λEe with decreasing soil water availability (e.g., Tan et al. 1978; Roberts et al. 1980; Kelliher et al. 1986). Black and Kelliher (1989) attributed this phenomenon to differences between overstory and understory coupling. They hypothesized that the well-coupled forest overstories, with Ω approaching 0.0, have strong stomatal control of λE and thus show reduced λE under soil or atmospheric drought conditions that induce stomatal closure. In contrast, they postulate that understory vegetation in these stands tended to have higher Ω because of its short stature and exposure to low wind speeds, which resulted in λEu being more responsive to Rnet than to VPD (Jarvis and McNaughton 1986; Martin et al. 2001). We did not estimate Ω for our understory because of the additional complexities and assumptions associated with those calculations, but it does not appear that λEu λEe in our stand responds to drought in the same way as it does in the stands described in Black and Kelliher (1989). We hypothesize that the decline in λEu λEe from winter to summer in the current study resulted primarily from increased overstory λE associated with elevated LAI during the summer, rather than from any differences between overstory and understory response to θ caused by differences between overstory and understory Ω. The fact that overstory tree LAI was correlated with mean monthly midday λEe (r = 0.44, p = 0.0481) but not with λEu (r = 0.12, p = 0.6059) supports this hypothesis. The canopy water-holding capacity of 0.19 mm for the ACMF stand was much lower than the 0.42–0.50 mm for closed-canopy plantations (Liu 1998; Gholz and Clark 2002), reflecting the lower LAI of this stand. However, canopy interception was 20% of annual precipitation at the ACMF in both years, whereas interception was only 8% of precipitation for the plantation stands (Gholz and Clark 2002). The high interception but lower water-holding capacity for the natural stand is probably related to the much higher annual precipitation during the Gholz and Clark study. However, it is also likely that water intercepted during individual rain events evaporated more rapidly from the more exposed surfaces of the more open structured ACMF canopy. © 2005 NRC Canada 1578 Evapotranspiration has been estimated for numerous Florida flatwoods ecosystems using a range of techniques, including weighing lysimeters, mass balance models, sap flux gauges, soil moisture changes, and micrometeorology (Allen et al. 1982; Riekerk 1982, 1985, 1989; Golkin and Ewel 1984; Liu 1996; Liu et al. 1998; Gholz and Clark 2002). Allen et al. (1982) used the mass balance equation in computer simulations of ET for watersheds in south Florida that were at least 50% pine and estimated ET to be 70% of annual precipitation. Using weighing lysimeters, Riekerk (1982) found that annual ET accounted for 60%–90% of precipitation for young slash pine stands. Using a process based model, Liu (1996) estimated annual ET to be 783–876 mm·year–1 for 1991–1993 for nearby slash pine stands that had similar maximum canopy LAI values (2.6) and somewhat higher stocking (544 stems·ha–1) as compared with the ACMF forest. On a daily basis in the winter of 1997–1998, sap flux density measurements of a young managed flatwoods plantation revealed that for identical climate conditions, loblolly pine transpired 1.8 mm·day–1, while slash pine transpired 1.5 mm·day–1 (Martin 2000). Therefore, the ACMF annual ET estimates of 676–832 mm·year–1 and daily ET estimates of 1.3–2.7 mm·day–1 from this study compare well with previous pine flatwoods ET estimates made at various spatial and temporal scales and with very contrasting methodologies. Although previous research in pine flatwoods forests in this region has suggested that soil or atmospheric drought seldom impacts tree physiological function (Neary et al. 1990; McMurtrie et al. 1994; Teskey et al. 1994), the present study took place in two remarkably dry years, and the impacts on ecosystem function were evident. The effects of drought on energy partitioning were manifested in several ways. First, reduced canopy conductance in response to elevated VPD (presumably caused primarily by stomatal closure) resulted in moderation of the partitioning of radiation to λE as VPD increased, even when soils were relatively moist (Fig. 5). This response was modified as soils dried below 6% volumetric water content, with λE Rnet actually declining as VPD increased above 1.75 kPa (Fig. 5). This was a consequence of both an overall decline in gc with drying soil and a probable increase in the sensitivity of gc to VPD (Ewers et al. 2001). Finally, monthly patterns of radiation-normalized λE appeared to be sensitive to short-term reductions in LAI associated with drought-induced needlefall pulses (Fig. 2). While these responses are not unusual in terms of general forest drought effects (e.g., Irvine et al. 1998; Ewers et al. 2001), they are out of the ordinary for these flatwoods ecosystems, where soil water availability is seldom an issue because of the presence of a watertable within 1 m of the surface (Abrahamson and Hartnett 1990; Teskey et al. 1994). It is likely that the receding watertable during the course of this study (Fig. 2) induced or enhanced physiological responses to VPD and drying surface soil in this forest. The Ω decoupling coefficient varied monthly and appeared to be closely tied to the depth of the watertable (Fig. 2). The decline in gc as soils dried was a primary driver of monthly variation in Ω (data not shown). Ω is essentially an index of the ratio gc/ga (Jarvis and McNaughton 1986). Since ga had no discernible seasonal trends, variation in gc was the main determinant of changes in Ω in our study. Despite noticeable monthly variation in Ω, even the highest values were well Can. J. For. Res. Vol. 35, 2005 within the ranges reported for conifer forests, where low gc and high ga tends to be the norm, leading to strong control of ET by gc and VPD (Jarvis and McNaughton 1986; Martin et al. 2001). While variation in annual precipitation precludes strict comparison of the magnitude of annual ET between the ACMF stand and surrounding plantations, the proportions of precipitation consumed by the pine stands in the region are similar: 80%–86% for the ACMF and the 2-year-old stand, and 92%–113% for the closed-canopy plantations (Gholz and Clark 2002). In other words, annual ET for the various Florida pine systems seems to fluctuate with precipitation, while maintaining a relatively constant ET/P. This contrasts with a ponderosa pine forest in Oregon with annual precipitation of 595 and 188 mm during 2 years, which maintained similar annual ET of 430 and 400 mm, but where the ET/P ratios were 0.72 and 2.12 (Anthoni et al. 1999). One interpretation is that the Florida pines avoid drought by closing their stomata to regulate water loss, whereas the ponderosa pine tolerate droughts and maintain ET by accessing alternative water sources with deep taproots (Anthoni et al. 1999). In conclusion, this naturally regenerated pine flatwoods ecosystem partitions net radiation into λE and H in proportions similar to those of other conifer forests. Local silvicultural practices affect canopy structure and hence the transfer of momentum from the atmosphere and albedo, and the partitioning of energy between λE and H. However, silviculture and stand age do not seem to affect annual ET as much as do fluctuations in annual precipitation. Clearly, the understory in the open-canopy stand in the present study plays a more significant role in ecosystem hydrology because of the greater understory LAI and enhanced penetration of radiation through the overstory canopy. For this reason, a priority for future research should be a better understanding of the hydrologic consequences of the frequent, low-intensity fires that characterize flatwoods managed under nonintensive silvicultural regimes, and which result in the removal and regrowth of the understory during each burning cycle. Acknowledgements This research was supported by the Biological and Environmental Research Program, US Department of Energy, through the Southeast Regional Center of the National Institute for Global Environmental Change; the NASA Land Cover and Land Use Change Program; and the University of Florida School of Forest Resources and Conservation. We also thank Ryan Atwood, Jose Luis Hierro, Jennifer Staiger, Jennifer Jacobs, and Julie Graves for their contributions to this work. References Abrahamson, W.G., and Hartnett, D.C. 1990. Pine flatwoods and dry prairies. In Ecosystems of Florida. Edited by R.L. Myers and J.J. Ewel. University of Central Florida Press, Orlando, Fla. pp. 103–149. Allen, E.R., and Gholz, H.L. 1996. Air quality and atmospheric deposition in southern U.S. forests. In Impact of air pollutants in southern pine forests. Ecological studies 118. Edited by Susan Fox and R.A. Mickler. Springer-Verlag New York, Inc., New York. pp. 83–170. © 2005 NRC Canada Powell et al. Allen, L.H., Jr., Knisel, W.G., Jr., and Yates, P. 1982. Evapotranspiration, rainfall, and water yield in South Florida research watersheds. Proc. Soil Crop Sci. Soc. Fl. 4: 127–128. Anthoni, P.M., Law, B.E., and Unsworth, M.H. 1999. Carbon and water vapor exchange of an open-canopied ponderosa pine ecosystem. Agric. For. Meteorol. 95: 151–168. Baldocchi, D.D., and Vogel, C.A. 1996. Energy and CO2 flux densities above and below a temperate broad-leaf forest and a boreal pine forest. Tree Physiol. 16: 5–16. Baldocchi, D.D., and Vogel, C.A. 1997. Seasonal variation of energy and water vapor exchange rates above and below a boreal jack pine stand. J. Geophys. Res. 102: 28 939 – 28 951. Baldocchi, D.D., Valentini, R., Running, S., Oechel, W., and Dahlman, R. 1996. Strategies for measuring and modelling carbon dioxide and water vapor fluxes over the terrestrial ecosystems. Global Change Biol. 2: 159–168. Berbigier, P., Bonnefond, J.M., Loustau, D., Ferreira, M.I., David, J.S., and Pereira, J.S. 1996. Transpiration of a 64-year old maritime pine stand in Portugal. Oecologia, 107: 43–52. Black, T.A., and Kelliher, F.M. 1989. Processes controlling understorey evapotranspiration. Phil. Trans. R. Soc. Lond. B, 324: 207–231. Brown, M.J. 1996. Forest statistics for Florida, 1995. USDA For. Serv. Res. Bull. SRS-6. Chen E., and Gerber, J.F. 1990. Climate. In Ecosystems of Florida. Edited by R.L. Myers and J.J. Ewel. University of Central Florida Press, Orlando, Fla. pp. 11–34. Clark, K.L., Gholz, H.L., Moncrieff, J.B., Cropley, F., and Loescher, H.W. 1999. Environmental controls over net carbon dioxide from contrasting Florida ecosystems. Ecol. Appl. 9: 936–948. Clark, K.L., Gholz, H.L., and Castro, M.S. 2004. Carbon dynamics along a chronosequence of slash pine plantations in N. Florida. Ecol. Appl. 14: 1154–1171. Dalla-Tea, F., and Jokela, E.J. 1991. Needlefall, canopy light interception, and productivity of young intensively managed slash and loblolly pine stands. For. Sci. 37: 1298–1313. Dougherty, P.M., Hennessey, T.C., Zarnoch, S.J., Stenberg, P.T., Holeman, R.T., and Wittwer, R.F. 1995. Effects of stand development and weather on monthly leaf and biomass dynamics of a loblolly (Pinus taeda) stand. For. Ecol. Manage. 72: 213–227. Ewers, B.E., Oren, R., Phillips, N., Stromgren, M., and Linder, S. 2001. Mean canopy stomatal conductance responses to water and nutrient availabilities in Picea abies and Pinus taeda. Tree Physiol. 21: 841–850. Gaston, L., Nkedi-Kizza, P., Sawka, G., and Rao, P.S.C. 1990. Spatial variability of morphological properties at a Florida flatwoods site. Soil Sci. Soc. Amer. J. 54: 527–533. Gholz, H.L., and Clark, K.L. 2002. Energy exchange across a chronosequence of slash pine forest in N. Florida. Agric. For. Meteorol. 112: 87–102. Gholz, H.L., Vogel, S.A., Cropper, W.P., Jr., McKelvey, K., Ewel, K.C., Teskey, R.O., and Curran, P.J. 1991. Dynamics of canopy structure and light interception in Pinus elliottii stands, north Florida. Ecol. Monogr. 61: 33–51. Gholz, H.L., Guerin, D.N., and Cropper, W.P. 1999. Phenology and productivity of saw palmetto (Serenoa repens) in a north Florida slash pine plantation. Can. J. For. Res. 29: 1248–1253. Golkin, K.R., and Ewel, K.C. 1984. A computer simulation of the carbon, phosphorus, and hydrologic cycles of pine flatwoods ecosystems. Ecol. Model. 24: 113–136. Goulden, M.L., Munger, J.W., Song-Miao Fan, Daube, B.C., and Wofsy, S.C. 1996. Measurements of carbon sequestration by 1579 long-term eddy covariance: methods and a critical evaluation of accuracy. Global Change Biol. 2: 169–182. Irvine, J., Perks, M.P., Magnani, F., and Grace, J. 1998. The response of Pinus sylvestris to drought: stomatal control of transpiration and hydraulic conductance. Tree Physiol. 18: 393–402. Jarvis, P.G. 1976. The interpretation of the variations in leaf water potential and stomatal conductance found in canopies in the field. Phil. Trans. Roy. Soc. Lon. B, 273: 593–610. Jarvis, P.G., and McNaughton, K.G. 1986. Stomatal control of transpiration: scaling up from leaf to region. Adv. Ecol. Res. 15: 1–49. Jarvis, P.G., James, G.B., and Landsberg, J.J. 1976. Coniferous forest. In Vegetation and the atmosphere. Vol. 2. Academic Press, London. pp. 171–240. Katul, G.G., Hsieh, C.I., Bowling, D., Clark, K., Shurpali, N., Turnipseed, A., Albertson, J., Tu, K., Hollinger, D., Evans, R., Orfferle, B., Anderson, D., Ellsworth, D., Vogel, C., and Oren, R. 1999. Spatial variability of turbulent fluxes in the roughness sublayer of an even-aged pine forest. Boundary-Layer Meteorol. 93: 1–28. Kelliher, F.M., Black, T.A., and Price, D.T. 1986. Estimating the effects of understorey removal from a Douglas-fir forest using a two-layer canopy evapotranspiration model. Water Resour. Res. 22: 1891–1899. Kelliher, F.M., Leuning, R., and Schulze, E.-D. 1993. Evaporation and canopy characteristics of coniferous forests and grasslands. Oecologia, 95: 153–163. Kelliher, F.M., Lloyd, J., Arneth, A., Byers, J.N., McSeveny, T.M., Milukova, I., Grigoriev, S., Panfyorov, M., Sogatchev, A., Varlargin, A., Ziegler, W., Bauer, G., and Schulze, E.D. 1998. Evaporation from a central Siberian pine forest. J. Hydrol. 205: 279–296. Law, B.E., Williams, M., Anthoni, P.M., Baldocchi, D.D., and Unsworth, M.H. 2000. Measuring and modelling seasonal variation of carbon dioxide and water vapour exchange of a Pinus ponderosa forest subject to soil water deficit. Global Change Biol. 6: 613–630. Liu, S. 1996. Evapotranspiration from cypress (Taxodium ascendens) wetlands and slash pine (Pinus elliottii) uplands in North-Central Florida. Ph.D. thesis, University of Florida, Gainesville, Fla. Liu, S. 1998. Estimation of rainfall storage capacity in the canopies of cypress wetlands and slash pine uplands in North-Central Florida. J. Hydrol. 207: 32–41. Liu, S., Riekerk, H., and Gholz, H.L. 1998. Simulation of evapotranspiration from Florida pine flatwoods. Ecol. Model. 114: 19–34. Martin, T.A. 2000. Winter season tree sap flow and stand transpiration in an intensively-managed loblolly and slash pine plantation. J. Sustain. For. 10: 155–163. Martin, T.A., and Jokela, E.J. 2004. Developmental patterns and nutrition impact radiation use efficiency components in southern pine stands. Ecol. Appl. 14: 1839–1854. Martin, T.A., Brown, K.J., Cermak, J., Ceulemans, R., Kucera, J., Meinzer, F.C., Rombold, J.S., Sprugel, D.G., and Hinckley, T.M. 1997. Crown conductance and tree and stand transpiration in a second-growth Abies amabilis forest. Can. J. For. Res. 27: 797– 808. Martin, T.A., Brown, K.J., Kucera, J., Meinzer, F.C., Sprugel, D.G., and Hinckley, T.M. 2001. Control of transpiration in a 220-yearold Abies amabilis forest. For. Ecol. Manage. 152: 211–224. McMurtrie, R.E., Gholz, H.L., Linder, S., and Gower, S.T. 1994. Climatic factors controlling the productivity of pine stands: a model-based analysis. Ecol. Bull. 43: 173–188. © 2005 NRC Canada 1580 Moncrieff, J.B., Massheder, J.M., Verhoeh, A., Elbers, J., Heusinkveld, B., Scott, H., de Bruin, H., Kaba, P., and Jarvis, P.J. 1997. A system to measure surface fluxes of energy, momentum and carbon dioxide. J. Hydrol. 188–189: 589–611. Monteith, J.L. 1965. Evaporation and environment. Symp. Soc. Exp. Biol. 19: 205–234. Myers, R.L., and Ewel, J.J. (Editors). 1990. Ecosystems of Florida. University of Central Florida Press, Orlando, Fla. Neary, D.G., Rockwood, D.L., Comerford, N.B., Swindel, B.F., and Cooksey, T.E. 1990. Importance of weed control, fertilization, irrigation and genetics in slash and loblolly pine early growth on poorly drained spodosols. For. Ecol. Manage. 30: 271–281. National Climatic Data Center. 2004. Climatic data, Gainesville Regional Airport, Gainesville, Florida. National Oceanic and Atmospheric Administration, Asheville, N.C. Owen, W. 2002. The history of native plant communities in the South. In Southern forest resource assessment. Edited by D.N. Wear and J.G. Greis. USDA For. Serv. Gen. Tech. Rep. GTR-SRS-53. pp. 47–61. Penman, H.L. 1948. Natural evaporation from open water, bare soil and grass. Proc. R. Soc. Lond. Ser. A193: 120–145. Riekerk, H. 1982. Pine tree evapotranspiration. Florida Water Resources Research Center Publication 62, University of Florida School of Forest Resources and Conservation, Gainesville, Fla. Riekerk, H. 1985. Lysimetric evaluation of pine forest evapotranspiration. In Forest–atmosphere interaction. Edited by B.A. Hutchison and B.B. Hicks. Reidel Publishing Company, Dordrecht, Netherlands. pp. 293–308. Riekerk, H. 1989. Influence of silvicultural practices on the hydrology of pine flatwoods in Florida. Water Resour. Res. 25: 713– 719. Roberts, J.M., Pymar, C.F., Wallace, J.S., and Pitman, R.M. 1980. Seasonal changes in leaf area, stomatal conductance and transpi- Can. J. For. Res. Vol. 35, 2005 ration from bracken below a forest canopy. J. Appl. Ecol. 17: 409–422. Roberts, J.M. 1983. Forest transpiration: a conservative hydrological process? J. Hydrol. 66: 133–141. Schmalzer, P.A., Hensley, M.A., and Dunlevy, C.A. 2001. Background characteristics of soils of Kennedy Space Center, Merritt Island, Florida: selected elements and physical properties. Fla. Sci. 64: 161–190. Schmid, H.P., Grimmond, C.S.B., Cropley, F., Offerle, B., and Su, H.B. 2000. Measurements of CO2 and energy fluxes over a mixed hardwood forest in the mid-western United States. Agric. For. Meteorol. 103: 357–374. Smith, W.B., Vissage, J.L., Darr, D.R., and Sheffield, R. 2001. Forest resources of the United States, 1997. USDA For. Serv. Gen. Tech. Rep. GTR-NC-219. Stevens, J. 1996. Applied multivariate statistics for the social sciences. Lawrence Erlbaum Associates Inc., Mahwag, N.J. Tan, C.S., Black, T.A., and Nnyamah, J.U. 1978. A simple diffusion model of transpiration applied to a thinned Douglas-fir stand. Ecology, 59: 1221–1229. Teskey, R.O., Gholz, H.L., and Cropper, W.P., Jr. 1994. Influence of climate and fertilization on net photosynthesis of mature slash pine. Tree Physiol. 14: 1215–1227. Twine, T.E., Kustas, W.P., and Norman, J.M. 2000. Correcting eddy-covariance flux underestimates over grassland. Agric. For. Meteorol. 103: 279–300. Wilson, K.B, Baldocchi, D.D., Aubinet, M., Berbigier, P., Bernhofer, C., Dolman, H., Falge, E., Field, C., Goldstein, A., Granier, A., Grelle, A., Halldor, T., Hollinger, D., Katul, G., Law, B.E., Lindroth, A., Meyers, T., Moncrieff, J., Monson, R., Oechel, W., Tenhunen, J., Valentini, R., Verma, S., Vesala, T., and Wofsy, S. 2002. Energy partitioning between latent and sensible heat flux during the warm season at FLUXNET sites. Water Resour. Res. 38: Art. 1294. © 2005 NRC Canada ...
View Full Document

Ask a homework question - tutors are online