Environ Monit Assess (2015)87: DOI 10.1007/s10661-015-4486-6

Assessing the influence of topography and canopy structure on Douglas fir throughfall with LiDAR and empirical data in the Santa Cruz mountains, USA K. T. Griffith & A. G. Ponette-González & L. M. Curran & K. C. Weathers

Received: 20 November 2014 / Accepted: 31 March 2015 # Springer International Publishing Switzerland 2015

Abstract Atmospheric inputs to forest ecosystems vary considerably over small spatial scales due to subtle ch an ges in re lief a nd v eg e t a t i o n s t r u c t u r e . Relationships between throughfall fluxes (ions that pass through the canopy in water), topographic and canopy characteristics derived from sub-meter resolution light detection and ranging (LiDAR), and field measurements were compared to test the potential utility of LiDAR in empirical models of atmospheric deposition. From October 2012 to May 2013, we measured bulk (primarily wet) deposition and sulfate–S, chloride (Cl−), and nitrate–N fluxes beneath eight clusters of Douglas fir trees differing in size and canopy exposure in the Santa Cruz Mountains, California. For all trees sampled, K. T. Griffith : A. G. Ponette-González (*) Department of Geography, University of North Texas, 1155 Union Circle #305279, Denton, TX 76203, USA e-mail: [email protected] L. M. Curran Woods Institute for the Environment, Stanford University, 473 Via Ortega, Stanford, CA 94305, USA L. M. Curran Department of Anthropology, Stanford University, 450 Serra Mall, Stanford, CA 94305, USA K. C. Weathers Cary Institute of Ecosystem Studies, PO Box AB, Millbrook, NY 12545-0129, USA Present Address: K. T. Griffith National Wetlands Research Center, US Geological Survey, 700 Cajundome Blvd, Lafayette, LA 70506, USA

LiDAR data were used to derive canopy surface height, tree height, slope, and canopy curvature, while tree height, diameter (DBH), and leaf area index were measured in the field. Wet season throughfall fluxes to Douglas fir clusters ranged from 1.4 to 3.8 kg S ha−1, 17–54 kg Cl− ha−1, and 0.2–4 kg N ha−1. Throughfall S and Cl− fluxes were highest under clusters with large trees at topographically exposed sites; net fluxes were 2–18-fold greater underneath exposed/large clusters than all other clusters. LiDAR indices of canopy curvature and height were positively correlated with net sulfate–S fluxes, indicating that small-scale canopy surface features captured by LiDAR influence fog and dry deposition. Although tree diameter was more strongly correlated with net sulfate–S throughfall flux, our data suggest that LiDAR data can be related to empirical measurements of throughfall fluxes to generate robust high-resolution models of atmospheric deposition. Keywords Atmospheric deposition . Forest . Montane . Nitrogen . Nutrients . Pollutants . Sulfur

Introduction Atmospheric deposition, the movement of nutrients and pollutants from the atmosphere to the Earth’s surface, is a critical input to forest ecosystems (Johnson and Lindberg 1992; Lovett 1994; Likens et al. 1996; Emmett et al. 1998; Fenn et al. 2010; Pardo et al. 2011; Weathers and Ponette-González 2011). Nutrients and pollutants are delivered to canopies in wet (rain,

270

Page 2 of 13

sleet, and snow), fog, and dry (particles and gases) forms; in many forest ecosystems, these pathways of deposition are the major source for macronutrients as well as salts, such as sulfur (S), nitrogen (N), and chloride (Cl) (Likens et al. 1981; Weathers et al. 2006; Schlesinger and Bernhardt 2013). Although S and N are essential nutrients for plant growth, in many temperate forests inputs of N, in particular, are much higher than what plants require (Weathers and Lovett 1998; Fenn et al. 1998; Driscoll et al. 2001; Aber et al. 2003; Liu et al. 2013). Montane forests, including those in California, are particularly vulnerable to elevated atmospheric deposition (Lovett and Kinsman 1990; Porter et al. 2005; Weathers et al. 2001, 2006; Ponette-González et al. 2010a; Fenn et al. 2003, 2010; Nanus et al. 2012). Plants in mountain ecosystems are often exposed to extreme conditions. Mountains are also wetter and windier than adjacent lowlands, and thus, total (wet+ fog+dry) atmospheric deposition is generally greater at high- than at low-elevation sites (Lovett and Kinsman 1990; Lovett et al. 1999; Weathers et al. 2000, 2006; Clow et al. 2015). In addition, tree species composition changes along elevational gradients in temperate regions, with needle-leaved trees frequently dominating at the highest elevations. Needle-leaved trees have greater leaf surface area (and, for evergreens, year round exposure) than broad-leaved trees, which contributes to enhanced dry and fog inputs (Beckett et al. 2000; Weathers et al. 2000; Erisman and Draaijers 2003; De Schrijver et al. 2007; Stankwitz et al. 2012). However, atmospheric deposition does not occur evenly across montane forest landscapes; at large spatial scales (100– 1000s of kilometers), inputs vary by proximity to emission source, topography, climate, and vegetation type and structure (National Atmospheric Deposition Program; Driscoll et al. 2001; Fenn et al. 2003). These same factors also change over short horizontal and vertical distances (50 cm DBH; n=4 clusters) size classes. Clusters of intermediate and large trees were visually identified as having either an exposed or a sheltered canopy. A cluster was identified as exposed when: (1) tree crowns in the cluster emerged above the crowns of neighboring trees, or (2) trees were located on steep upper slopes unsheltered by neighboring trees. Because aspect has been demonstrated to strongly influence atmospheric deposition rates (Weathers et al. 2006), all clusters were established on west-facing slopes to ensure that canopies were angled in the direction of the west and northwest prevailing winds. Field sampling Throughfall fluxes and bulk deposition Below these Douglas fir tree clusters, throughfall (water that drips from the forest canopy to the forest floor) was sampled to estimate SO42−–S; Cl−, and NO3−–N fluxes to the soil. Sulfate in throughfall was used as an index of total (wet+fog+dry) atmospheric deposition because of minimal uptake and leaching of this ion by plant canopies (Lindberg and Garten 1988). Nitrogen is a major limiting nutrient in temperate forest ecosystems that is actively taken up by and leached from plant canopies, and thus, total atmospheric deposition cannot be estimated for N using the canopy throughfall method (Weathers et al. 2006); however, N fluxes to the forest floor can be determined. In September 2012, 24 ion-exchange resin (IER) throughfall collectors were established at SDSF. Following the protocol of Simkin et al. (2004) and Weathers et al. (2006), each collector consisted of a plastic funnel (20 cm in diameter; 324 cm2) attached to a chromatograph column loaded with 20 ml of anionexchange resin (Dowex Monosphere 550A) mixed with double deionized (DDI) water. Ion-exchange resins are increasingly used in canopy throughfall studies because the stability of resin beads allows samples to remain in remote areas for weeks to a few months before columns exceed capacity and require replacement (Simkin et al. 2004). Throughfall water accumulates in the funnel and then passes through the column, where ions in throughfall adhere to charged resin surfaces.

270

Page 4 of 13

Environ Monit Assess072: 81 )5 02(

Fig. 1 a Map of study region in the Santa Cruz Mountains, California. Two research sites were selected within Soquel Demonstration State Forest (SDSF). b The lower helipad (HPAD; 37°05′22.74″ N, 121°53′ 26.75″ W; 418 m a.s.l; plots top left) and the herpetology study area (AMIB; 37°04′54.27″ N 121°53′09.67″ W; 593 m a.s.l.; plots lower right). c At each site, three throughfall collectors were placed beneath four separate clusters of Douglas fir trees, and two bulk collectors were installed in an adjacent clearing

Under each Douglas fir tree, a throughfall collector was established, totaling three collectors per cluster. Collectors were set 1 m aboveground on PVC pipes positioned midway between the bole and the canopy dripline (i.e., outer edge of tree crowns) to ensure that throughfall was collected and any bulk rainwater collection (see below) was avoided. To maintain sampling of Douglas fir throughfall, we avoided placing collectors under branches overlapping with neighboring trees. At each site, two identical bulk collectors were placed nearby (0, some combination of leaching (for non conservative indicators), fog, and dry deposition is inferred. When NTF < 0, canopy uptake is assumed. Sulfate–S in throughfall was used as an indicator of total (wet+ dry+fog) atmospheric deposition and SO42−–S in net throughfall as an indicator of dry plus fog deposition. For each sampling period, the throughfall and net throughfall fluxes per tree cluster were calculated by averaging the three individual collectors within a cluster. Values for the three sampling periods were then summed

270

Page 6 of 13

and expressed on a per season basis to obtain the wet season throughfall and net throughfall flux for each tree cluster. Relationships between topographic and vegetation structure variables measured in the field (e.g., DBH) as well as vegetation and topographic indices derived from LiDAR point cloud data (e.g., CHM) were compared to throughfall and net throughfall SO42−– S, Cl−, and NO3−–N fluxes for each sampling period separately and then for the entire wet season. A Shapiro–Wilk test was used to determine if the variables were normally distributed. Because most variables had skewed distributions, nonparametric Spearman’s rho (r) correlations were employed. Significant correlations were then examined using simple linear regression. Variables with skewed distributions were log-transformed to meet assumptions of normality. All analyses were performed using SPSS v. 20 (SPSS 2012) and JMP v. 11 (JMP 2013).

Results

Environ Monit Assess072: 81 )5 02(

Throughfall fluxes Throughfall SO42−–S fluxes––an indicator of total atmospheric deposition––measured under eight Douglas fir tree clusters ranged from 1.4 to 3.8 kg S ha−1 during the wet season (October 2012 to May 2013). Total SO42−–S deposition to exposed/large clusters was 1.7–2.7-fold greater than to exposed/intermediate, sheltered/large, and sheltered/intermediate clusters (Fig. 2). Fluxes of Cl− in throughfall ranged from 17.1 to 54.1 kg Cl− ha−1 per wet season and were similar to reported data from coastal temperate conifer sites in the northeastern USA and Japan (Aikawa et al. 2006; Weathers et al. 2006; Fig. 2). As with SO42−–S, Cl− fluxes were highest under exposed/large clusters. These clusters received 1.5–3.2fold more Cl− in throughfall compared to all other clusters. Nitrate–N throughfall fluxes ranged from 0.2 to 4 kg N ha−1 per wet season and differed between HPAD and AMIB sites within the study area (Fig. 2). During the sampling period, lower NO3−–N fluxes were measured at HPAD (0.2–1.4 kg N ha−1) than at AMIB (1.3–4 kg N ha−1 per wet season). At the AMIB site, the NO3−–N throughfall flux to the exposed/large cluster was 1.4–3fold higher than to exposed/intermediate, sheltered/large, and sheltered/intermediate clusters.

Rainfall and temperature From October 2012 to May 2013, total rainfall recorded by the rain gauge did not differ significantly from rainfall measured in SDSF during those same months between 1996 and 2012 (Mann–Whitney U; p=0.599). Total rainfall from October 1996 to May 2012 averaged 1040±26 mm (Linsley, Kraeger, and Associates, Ltd.), whereas total precipitation from October 2012 to May 2013 was 842 mm. The 2012–2013 wet season was relatively brief (35 rain days versus 44 rain days in 1996–2012) with an extended dry period. Total precipitation in November and December 2012, however, was nearly 2-fold greater than that recorded in 1996–2012, whereas rainfall from January through May 2013 was nearly 5-fold lower than in 1996–2012. At the study site, mean monthly temperature ranged from 7 to 20 °C. October through November 2012 (17.2–11.8 °C) temperatures were consistent with 1996–2012 averages (16.7–13.3 °C), whereas December 2012 through February 2013 (7.4–9.1 °C) and March 2013 through May 2013 (13.9–17.9 °C) temperatures were consistently cooler and warmer than respective 1996–2012 means (11.1–12.2 °C and 12.8–16.1 °C) (NOAA 2012).

Net throughfall fluxes Net SO42−–S throughfall fluxes across the eight clusters ranged from 0.1 to 2.4 kg S ha−1 per wet season, indicating high variability in dry and fog S inputs to these Douglas fir canopies. Exposed/large clusters had greater net throughfall fluxes (2.4–18.2-fold) than exposed/intermediate, sheltered/large, and sheltered/ intermediate clusters, suggesting much higher fog/dry deposition. Net throughfall Cl− fluxes were consistently positive and ranged from 3 to 39.7 kg Cl− ha−1 per wet season, suggesting dry and fog Cl− deposition to plus leaching from the forest canopy. These fluxes represented 17.8–73.4% of the total Cl− measured in throughfall and were 2–13-fold greater under exposed/large clusters than the other clusters (Fig. 2). Net throughfall fluxes for NO3−–N ranged from −1.2 to 2.7 kg N ha−1 per wet season. Negative values indicate higher rates of canopy uptake, whereas positive values indicate lower rates of canopy uptake relative to nitrate leaching and/or dry HNO3 deposition (Fig. 2). Net NO3−–N fluxes also varied between sites (Fig. 2). Negative net NO3−–N

Environ Monit Assess (2015)87:

Page 7 of 13 270

Fig. 2 Throughfall and net throughfall fluxes of SO42−–S, Cl−, and NO3−–N (kg/ha/wet season) collected from October 2012 to May 2013 under Douglas fir tree clusters at HPAD (dark gray) and AMIB (light gray) in Soquel Demonstration State Forest, Santa Cruz Mountains, California. Throughfall SO42−–S flux was used

as a conservative tracer of total (wet+fog+dry) atmospheric deposition and net SO42−–S throughfall as an index of fog and dry deposition. For Cl− and NO3−–N, when NTF

Assessing the influence of topography and canopy structure on Douglas fir throughfall with LiDAR and empirical data in the Santa Cruz mountains, USA.

Atmospheric inputs to forest ecosystems vary considerably over small spatial scales due to subtle changes in relief and vegetation structure. Relation...
5MB Sizes 0 Downloads 9 Views