A preliminary evaluation of throughfall sampling techniques in a mature coniferous forest

Rainfall, throughfall and stemflow were measured and canopy interception loss was derived for 14 rainfall events from June 22 to August 30, 2008 within a mature declining hybrid white spruce (Picea glauca (Moench) Voss × P. engelmannii Parry ex Engelm.)-subalpine fir (Abies lasiocarpa (Hook.) Nutt.)-lodgepole pine (Pinus contorta var. latifolia Dougl. ex Loud.) stand in south-central British Columbia, Canada. Stemflow was negligible during the study period, while, respectively, throughfall and canopy interception loss accounted for approximately 59.4% and 40.6% of the 50.1 mm of cumulative rainfall. Throughfall variability was assessed with three approaches involving roving and stationary wedge-type gauges, and stationary trough gauges. Throughfall exhibited large spatial variability with the coefficient of variability of study period throughfall sampled using 16 stationary trough gauges being 30.3%, while it was 38.0% and 28.7% for 32 stationary and 32 roving wedge gauges, respectively. Our analysis suggests that a roving gauge method is better than a stationary approach since the errors associated with event mean throughfalls are summed quadratically and a greater portion of the canopy area is sampled. Trough gauges were more efficient than wedge gauges; however, this efficiency was less than expected given their much larger sampling areas, suggesting that spatial autocorrelation lengths of throughfall may be longer than the trough systems. The spatial distribution of throughfall showed a high degree of temporal persistence throughout the study suggesting the existence of stable “wet” and “dry” inputs to the floors of these coniferous forests.