On estimating the exponent of power-law frequency distributions. Ecology 89, 905-912

Department of Biology and the Ecology Center, Utah State University, Logan, Utah 84322, USA.
Ecology (Impact Factor: 4.66). 05/2008; 89(4):905-12. DOI: 10.1890/07-1288.1
Source: PubMed


Power-law frequency distributions characterize a wide array of natural phenomena. In ecology, biology, and many physical and social sciences, the exponents of these power laws are estimated to draw inference about the processes underlying the phenomenon, to test theoretical models, and to scale up from local observations to global patterns. Therefore, it is essential that these exponents be estimated accurately. Unfortunately, the binning-based methods traditionally used in ecology and other disciplines perform quite poorly. Here we discuss more sophisticated methods for fitting these exponents based on cumulative distribution functions and maximum likelihood estimation. We illustrate their superior performance at estimating known exponents and provide details on how and when ecologists should use them. Our results confirm that maximum likelihood estimation outperforms other methods in both accuracy and precision. Because of the use of biased statistical methods for estimating the exponent, the conclusions of several recently published papers should be revisited.

Download full-text


Available from: Brian J. Enquist
  • Source
    • "Interestingly, the risk of misidentifying a (composite) correlated random walk as Lévy walk is strongly affected by the data sampling scheme and whether higher dimensional data is projected on a lower dimension (Plank and Codling 2009; Codling and Plank 2011). The debate on Lévy walks further concerns statistical methods that are used to detect Lévy walk behaviour in data (White et al 2008; Edwards 2011). Even the application of Lévy walks within optimal foraging theory as Lévy foraging hypothesis has been met with scepticism (Pyke 2015). "
    [Show abstract] [Hide abstract]
    ABSTRACT: When sampling animal movement paths, the frequency at which location measurements are attempted is a critical feature for data analysis. Important quantities derived from raw data, e.g. travel distance or sinuosity, can differ largely based on the temporal resolution of the data. Likewise, when movement models are fitted to data, parameter estimates have been demonstrated to vary with sampling rate. Thus, biological statements derived from such analyses can only be made with respect to the resolution of the underlying data, limiting extrapolation of results and comparison between studies. To address this problem, we investigate whether there are models that are robust against changes in temporal resolution. First, we propose a mathematically rigorous framework, in which we formally define robustness as a model property. We then use the framework for a thorough assessment of a range of basic random walk models, in which we also show how robustness relates to other probabilistic concepts. While we found robustness to be a strong condition met by few models only, we suggest a new method to extend models so as to make them robust. Our framework provides a new systematic, mathematically founded approach to the question if, and how, sampling rate of movement paths affects statistical inference.
    Full-text · Article · Feb 2016 · Journal of Mathematical Biology
  • Source
    • "For each size distribution, the slope was calculated using three different interval rounding procedures (upper unit, lower unit, and 0.5 unit rounding) and averaged to avoid rounding (Bobe and Behrensmeyer 2004; Kelt and Meyer 2009). Previous work showed that distribution analyses that use the number of events found inside of log-transformed intervals (like the number of species found inside a given bin size interval) are biased because larger intervals tend to contain more events (e. g. species) just by chance (Bonnet et al. 2001; Edwards 2008; White et al. 2008 ). In order to correct for this chance effect, the number of species found on each interval was divided by the bin interval width (Bonnet et al. 2001). "
    [Show abstract] [Hide abstract]
    ABSTRACT: Our objective was to estimate and analyze the body-size distribution parameters of terrestrial mammal assemblages at different spatial scales, and to determine whether these parameters are controlled by local ecological processes or by lar- ger-scale ones. Based on 93 local assemblages, plus the complete mammal assemblage from three continents (Africa, North, and South America), we esti- mated three key distribution parameters (diversity/size slope, skewness, and modal size) and compared the values to those expected if size distributions are mainly controlled by local interactions. Mammal diversity decreased much fas- ter as body size increased than predicted by fractal niche theory, both at conti- nental and at local scales, with continental distributions showing steeper slopes than the localities within them. South America showed a steeper slope (after controlling for species diversity), compared to Africa and North America, at local and continental scales. We also found that skewness and modal body size can show strikingly different correlations with predictor variables, such as spe- cies richness and median size, depending on the use of untransformed versus log-transformed data, due to changes in the distribution density generated by log-transformation. The main differences in slope, skewness, and modal size between local and continental scales appear to arise from the same biogeo- graphical process, where small-sized species increase in diversity much faster (due to higher spatial turnover rates) than large-sized species. This process, which can operate even in the absence of competitive saturation at local scales, generates continental assemblages with steeper slopes, smaller modal sizes, and higher right skewness (toward small-sized species) compared to local communi- ties. In addition, historical factors can also affect the size distribution slopes, which are significantly steeper, in South American mammal assemblages (prob- ably due to stronger megafauna extinction events in South America) than those in North America and Africa.
    Full-text · Article · Jan 2016 · Ecology and Evolution
  • Source
    • "The intercept can indicate total or standing stock biomass (Sprules and Munawar 1986; Boudreau and Dickie 1992) while the slope represents ecosystem structure, i.e., the relative contribution of small versus large organisms, which result from underlying energy transfer efficiency and predator-prey size ratios (Sprules and Munawar 1986; Gómez-Canchong et al. 2013; Yurista et al. 2014). Normalized biomass size spectrum slopes are expected to be around - 1 to -2 depending on whether biomass or number of organisms are used represent density (Andersen and Beyer 2006; White et al. 2008; Basset and Sabetta 2009) (Fig. 1). The expected slope from NLSS whereby body size is characterized by length as opposed to weight, is -6 based on the length-weight relationship; W=L 3 where W is weight and L is length (Fig. 1). "

    Full-text · Article · Nov 2015 · Canadian Journal of Fisheries and Aquatic Sciences
Show more