Manipulations during the 2007–2008 growing seasons consisted of five
levels of precipitation (PPT) relative to ambient: −80, −50%, control,
+50, +80%, with 12 replicates per treatment. Plots were 2.5 × 2.5 m,
and each was centered around a single mature mesquite shrub that was
similar in size to other mesquite on this ecological site (mean
height 0.5 m, canopy diameter 1.1 m). Shrubs were surrounded by black
grama . Reductions in PPT were achieved via passive rainout shelters;
supplementation was achieved via a sprinkler irrigation system.
Desired levels of rainfall reduction (−80 and −50%) relative to
ambient were achieved by modifying the number of transparent V-shaped
acrylic strips per shelter [Acrylite FF; CYRO Industries, Parsippany,
NJ; see Yahdjian and Sala 2002 for a detailed description of shelter
design and their minimal effect on the microenvironment]. A PVC-pipe
irrigation system was used at the enhanced PPT plots, which were
irrigated the day after PPT events >2 mm, with an amount of
water equal to 50 or 80% of the event. Supplementation occurred during
the spring and fall of 2007 and 2008 using PPT that was captured
nearby and stored. Control plots received ambient PPT and had neither
shelters nor irrigation systems.
Soil volumetric water content (VWC) was quantified simultaneously with
ecophysiological measurements. VWC was measured in six plots per
treatment at depths of 0–5 and 30–50 cm using EC-5 and EC-20 soil
moisture sensors (Decagon Devices, Pullman, WA), respectively. Probes
were calibrated for soils at the site following the protocol developed
by the manufacturer (r-squared of 0.98 and 0.96 for EC-5 and EC-20
probes, respectively).
