Short communication Reducing sampler error in soil research Andrew Kulmatiski a, * , Karen H. Beard b a Department of Biology and the Ecology Center, Utah State University, Logan, UT 84322-5305, USA b Department of Forest, Range and Wildlife Sciences and the Ecology Center, Utah State University, Logan, UT 84322-5230, USA Received 21 March 2003; received in revised form 2 September 2003; accepted 6 October 2003 Abstract Most soil sampling techniques are calibrated with standards and against each other, but the error introduced by the individuals performing each technique (sampler error) is often not addressed. We determined that sampler error explains a minimum of 5.5% of the variation in soil bulk density when using a composite coring technique. This source of error is of concern because it could easily obscure the small, but significant differences anticipated with long-term (decadal) research. We suggest three methods for increasing sampling accuracy across spatial and temporal treatments: (1) use . 5 individuals to sample each treatment; (2) consider both the concentration and pool size of soil properties (e.g. nitrate, root mass, etc.); and (3) include measurements of sampler error. q 2003 Elsevier Ltd. All rights reserved. Keywords: Bulk density; Coring technique; Sampler error; Sampling; Sampling design Soil sampling is used by soil ecologists to measure root, fungal, bacterial, and invertebrate densities; by plant ecologists to determine the size and activity of plant nutrient pools; by climate modelers to determine biosphere/atmosphere exchange; and by environmental chemists to monitor pollution loading (Wagner et al., 2001). Across these disciplines, estimates of soil bulk density are needed to extrapolate measurements from soil samples to the landscape scale (Vincent and Chadwick, 1994). Yet, microsite heterogeneity, variability in sampling approaches, and variability among individuals performing the sampling (henceforth, the samplers) all increase the uncertainty around estimates of soil properties (Johnson et al., 1990). Variability associated with microsite heterogeneity can be addressed with spatially explicit sampling designs (Klironomos et al., 1999; Legendre et al., 2002; Stein and Ettema, 2003), and appropriate sample volumes or replicate numbers (Vincent and Chadwick, 1994). Variability between sampling techniques has partially been addressed by calibrating different techniques against each other (Lichter and Costello, 1994; Page-Dumroese et al., 1999; Miller et al., 2001; Kulmatiski et al., 2003). However, we were unable to find any published estimates of sampling error attributable to variation between samplers, even though variation in sampling technique between individual samplers may introduce large errors to estimates of soil properties. A wide array of techniques have been developed and employed to determine soil bulk density but, perhaps the most common technique remains the composite coring technique (Boone et al., 1999). This technique has been suggested for use in long-term ecological research as well as for use in the national Forest Health Monitoring program and is cited in the soil sampling guidelines used in European countries (Theocharopoulos et al., 2001; Palmer et al., 2002). It was our objective to estimate the error introduced to measurements of volumetrically sampled soil masses by individual soil samplers and to use this information to suggest a sampling design for soil research that acknowl- edges and minimizes sampler error. Sampling was performed from 15 to 31 July 2001 in two locations separated by 8 km in Methow Valley, Washington, USA (48837 0 N, 107810 0 W). Sampling was conducted in shrub-steppe vegetation on the Newbon soil series, a well-drained Typic Haploxeroll derived from glacial till. Two-hundred quadrats (1 m £ 1 m) were placed at 1 m intervals along transects of varying length which were fitted between large shrubs. Transects were placed at least 2 m and not more than 200 m apart in each field. Each sampler was assigned to sample 50 randomly selected plots from each of two field sites. 0038-0717/$ - see front matter q 2003 Elsevier Ltd. All rights reserved. doi:10.1016/j.soilbio.2003.10.004 Soil Biology & Biochemistry 36 (2004) 383–385 www.elsevier.com/locate/soilbio * Corresponding author. Tel.: þ 1-435-770-9646; fax: þ1-435-797-3796. E-mail address: andrew@biology.usu.edu (A. Kulmatiski).