Measuring Fractional Cover and Leaf Area Index in Arid Ecosystems: Digital Camera, Radiation Transmittance, and Laser Altimetry Methods

Field measurement of shrubland ecological properties is should be guided by consideration of the amount of time and resources required to obtain measurements of the deimportant for both site monitoring and validation of remote sensing information. During the May 1997 NASA Earth sired variables. Our results suggest that the ADC is both Observing System Jornada Prototype Validation Exercise, efficient and accurate for long-term or large-scale monitorwe calculated plot-level plant area index, leaf area index, ing of arid ecosystems. Elsevier Science Inc., 2000 total fractional cover, and green fractional cover with data from four instruments: (1) a Dycam Agricultural Digital Camera (ADC), (2) a LI-COR LAI-2000 plant canopy INTRODUCTION analyzer, (3) a Decagon sunfleck Ceptometer, and (4) a Shrublands exist in hot, dry areas where high evaporative laser altimeter. Estimates from the LAI-2000 and Ceptodemand greatly exceeds unpredictable and sparse precipimeter were very similar (plant area index 0.3, leaf area tation (Evenari, 1985). Although estimates vary widely index 0.22, total fractional cover 0.19, green fractional (Townshend et al., 1991), pure shrublands cover approxicover 0.14), while the ADC produced values 5% to 10% mately 9% of the Earth’s vegetated surface (Waring and higher. Laser altimeter values, depending on the height Running, 1998). Within the past century, many arid to cutoff used to establish total fractional cover, were either semiarid areas of the United States have experienced drahigher or lower than the other instruments’ values: a 10-cm matic shrub increases, usually at the expense of native cutoff produced values ~80% higher, while a 20-cm cutoff grasses (Smith et al., 1997). While some shrub expansion produced values ~30% lower. The LAI-2000 and Ceptometer may be related to persistent drought (Herbel et al., 1972), are designed to operate in homogenous canopies, not the evidence suggests that overgrazing and fire suppression sparse and irregular vegetation found at Jornada. Thus, are more important causes (Archer et al., 1995; Bryant et these instruments were primarily useful for relative withinal., 1990; Grover and Musick, 1990). Such conversions can site plant area index monitoring. Calculation of some pabe detrimental to pastoral societies directly dependent on rameters required destructive sampling, a relatively slow grassland extent and productivity. High shrub cover may and labor-intensive activity that limits spatial and temporal also have beneficial effects, such as increasing runoff water applicability. Validation/monitoring campaigns therefore for irrigation (Skarpe, 1990) or accelerating aquifer recharge (Leduc et al., 1997). Thus, depending on local priorities, increased or decreased shrub populations may be * Numerical Terradynamic Simulation Group, School of Forestry, University of Montana, Missoula, MO desired. Regardless of the goal, accurate monitoring of † Cooperative Institute for Research in Environmental Sciences, shrubland extent and vigor is important for natural resource University of Colorado, Boulder, CO managers and for the people they serve. ‡ NASA Goddard Space Flight Center, Code 923, Biospheric Sciences Branch Satellite remote sensing provides the only technically Address correspondence to Michael A. White, NTSG/School of consistent and temporally regular means of monitoring Forestry, University of Montana, Missoula, MT 59812, USA. E-mail: mike@ shrublands over large areas. In shrublands, remote sensing ntsg.umt.edu Received 12 June 1998; revised 5 January 1999. is hampered by a high proportion of bare soil, clump shad-