Aboveground biomass in three Sonoran Desert communities: Variability within and among sites using replicated plot harvesting

Abstract

Total aboveground biomass (TAGB) values are reported for the first time for the Sonoran Desert. Harvesting of replicated plots in three sites differing in geomorphology and vegetation structure allowed the statistical measurement of spatial variability within and among sites. Linear, power log-transformed, and power non-linear regressions were used to relate TAGB with plant metrics. Canopy volume explained the largest proportion of TAGB variance (r² = 0.74–0.94) in the three models. All models were highly significant, but the non-linear was more robust, had a better distribution of residuals, and did not require data back-transforming, allowing the accurate estimation of biomass at the plot level using simple measurements of vegetation. TAGB ranged from 6.99 Mg ha⁻¹ (Plains) to 29.24 Mg ha⁻¹ (Arroyos) in the desertscrub, and was intermediate in the thornscrub (Hillsides: 13.03 Mg ha⁻¹). Within-site variability of TAGB was highest in the Plains and lowest in the Hillsides. The weighted mean AGB was 16 Mg ha⁻¹. Assuming a 1:1 root:shoot ratio, average total biomass could be 32 Mg ha⁻¹. This remarkable value sets the Sonoran Desert as a potential reservoir of up to 4.4% of the total world deserts plant biomass in less than 1% of the deserts’ global area.

Introduction

Accurate quantifications of carbon storage in natural vegetation are important to determine the magnitude of biomass and carbon stocks in terrestrial ecosystems and to improve our understanding of how ecosystems might respond to climate change (Lioubimtseva and Adams, 2004, Litton and Kauffman, 2008). They are also important to properly assess the impacts of land-use/land-cover change on the terrestrial carbon balance (Jaramillo et al., 2003, Bonino, 2006, Piao et al., 2009). Different methods have been used to determine the amount of aboveground biomass (AGB) and carbon stocks in terrestrial vegetation. Most studies have relied on generalized dimension analysis (sensu Allaby, 1998), an indirect method consisting of establishing size-biomass allometric relationships of selected species. By this technique, the size of the plants prior to destructive sampling (harvesting and weighing) is measured, and later, specific plant size-biomass relationships are established (Baskerville, 1972, Li and Xiao, 2007, Litton and Kauffman, 2008, Návar, 2009). This method is easy to apply in plant communities composed of one or a few codominant species (Patten, 1978, Rundel et al., 1982, Návar et al., 2002, Northup et al., 2005). However, in more diverse and structurally complex communities, a larger number of samples are often needed to account for the higher spatial variability in biomass. Direct methods of biomass measurement involving the harvesting of whole plots allow the validation of other methods such as dimension analysis models or remote sensing techniques aimed to generalize biomass estimates (Houghton et al., 2009).

A widespread perception of drylands is that the total amount of biomass is small and evenly distributed. This idea is based on the assumption that vegetation cover in arid and semi-arid ecosystems –that represent a sizable portion of the world’s land area (29.8%; Lal, 2004)– is sparse compared to the more humid regions. However, in many desert areas, biomass and productivity can be considerably high and variable, even at small spatial scales (hundreds of metres; Huenneke and Schlesinger, 2006). This variation in the distribution of biomass is proximally affected by differences in soil water availability (Noy-Meir, 1973, Evenari and Noy-Meir, 1985, Búrquez et al., 1999, López-Portillo and Montaña, 1999, Whitford, 2002, Rango et al., 2006), although, at smaller scales, it is known that biotic influences play a major role in determining vegetation dynamics (Valiente-Banuet and Ezcurra, 1991, Búrquez and Quintana, 1994, Callaway, 1995, Bisigato et al., 2009).

In the Sonoran Desert, spatial heterogeneity in vegetation structure –even among adjacent sites– has been related to strong differences in primary productivity (Maya and Arriaga, 1996, Martínez-Yrízar et al., 1999, Sponseller and Fisher, 2006), microbial potential of soils (Núñez et al., 2001) and decomposition rates (Arriaga and Maya, 2007, Martínez-Yrízar et al., 2007). Plant biomass is also likely to exhibit strong spatial variability, but its quantification has been a neglected aspect of ecosystem studies in this region.

The aims of the present study were to measure desert TAGB and its spatial variability applying the direct method of plot harvesting. Specific objectives were: (1) to quantify TAGB and carbon stocks in three extensive Sonoran Desert sites strongly differentiated by vegetation structure, topography and the presence of small drainage systems, (2) to develop the best multispecies regression models for the estimation of TAGB harvesting replicated plots, and (3) to briefly discuss the relevance of the Sonoran Desert carbon stocks on local and global scenarios of carbon balance.