Population dynamics and orientation of the central Andean cushion Azorella compacta

12 Azorella compacta ( llareta ; Apiaceae), a large woody cushion plant, is a remarkable 13 species forming dense cushions and characterizing the high elevation rocky slopes of the central 14 Andean Altiplano. Field studies of an elevational gradient of A. compacta within Lauca National 15 Park in northern Chile found a reverse J-shape distribution of size classes of individuals with 16 abundant small plants at all elevations. A new elevational limit for A. compacta was found at 17 5250 m. A series of cushions marked 14 years earlier showed either slight shrinkage or small 18 degrees of growth up to 2.2 cm yr -1 . Despite their irregularity in growth, cushions of A. compacta 19 show a strong orientation, centered on a north-facing aspect and angle of about 20 o from 20 horizontal. This exposure to maximize solar irradiance closely matches previous observations of 21 a population favoring north-facing slopes at a similar angle. Populations of A. compacta appear 22 to be stable, or even expanding, with young plants abundant and recolonization of disturbed 23 habitats.


INTRODUCTION
Azorella compacta (Apiaceae), a large woody cushion plant, forms an iconic species of the 33 Altiplano Plateau of northern Chile, Bolivia, Argentina and Peru (Kleier & Rundel 2004). 34 Known locally as llareta, it forms broad irregular cushions that commonly reach diameters of 3-35 4 m, or much more, on rocky slopes at high elevations. Its range extends across the Altiplano 36 Plateau of the south-central Andes from southern Peru through western Bolivia and into the 37 northeastern Chile and north-western Argentina (Martinez 1993). The species only rarely occurs 38 below 4000 m and an upper elevational limit of 5200 m has been reported (Halloy 2002), making 39 it one of the highest occurring woody plant species in the world. continuous cushion up to 10 m or more across. Previous research on A. compacta has described 47 microhabitat selection, population structure, germination rate, and growth rate of the plants 48 (Kleier & Rundel 2004), as well as the ecophysiology and energy balance of the A. compacta 49 cushions (Kleier & Rundel 2009). There have also been concerns about conservation of the 50 species due to past major harvesting of llareta for fuel in the early and mid-20 th century, and A. 51 compacta continues to be classified as a "data deficient" species (IUCN 2012). 52 The present research continues a long-term study of A. compacta begun in 1998, and expands    The tape measure was allowed to follow the surface of the plant to account for irregular planar   land. In 1998, park staff indicated that the proposed plots would be located on public land.

109
However, in 2012, we found several painted messages denoting the area as private property.

110
While removal of the tags lessened the value of data gained from the plots, it indicates 111 underlying desires of local Amayara populations to own and protect the park's resources. 112 Notably, we found no signs of recent llareta harvesting in this area.

113
For the nine remaining tagged plants, we measured length and width in orthogonal axes 114 across the apex of the cushion, perimeter, and height, which was determined from the apex of the 115 cushion to the nearest western edge. We also noted any dieback (increase in dead tissue) and the 116 presence of flowers or fruits. that transect line and the angle from horizontal was measured (Fig. 2). The same field assistant 137 was employed for all measurements to avoid changes in bias between individual measurements.

138
The declination from magnetic north of 5.33° W was determined for latitude 18°12'6.70" S,   Table 1 and Fig. 4. A bootstrap analysis to determine a 95% confidence interval of growth rate 160 around the mean and the median based on 10,000 replicates found a range from -2.86 cm yr -1 to 161 0.55 cm yr -1 . The negative growth rates resulted from a decrease in the individual perimeter over 162 the sample interval. When the median was used for bootstrapping, the 95% confidence intervals 163 around perimeter growth ranged -2.55 cm yr -1 to 2.09 cm yr -1 . These changes in perimeter 164 corresponded to a radial growth rate of -0.5 to 0.4 cm yr -1 . 165 Despite these slow rates of mean growth, we also found that individual A. compacta can 166 grow significantly more quickly under some conditions. As an example of rapid growth, we