SPATIAL VARIATION IN THE FORCE REQUIRED TO INITIATE ROCK MOVEMENT IN 4 UPLAND STREAMS - IMPLICATIONS FOR ESTIMATING DISTURBANCE FREQUENCIES

Lotic models of disturbance generated by floods and spates suffer from 2 main shortcomings: a lack of knowledge regarding the appropriate sp atial scale at which to apply models and a poor understanding of the r elationship between discharge sizes and actual disturbance frequencies and intensities. Here, we examine the spatial variability in the forc es needed to shift rocks and the utility of hydraulic equations that p redict critical shear stresses (tau(c)), which are sometimes used to i nfer disturbance frequencies in streams. We used spring balances to me asure directly the forces needed (F-c) to move rocks up and out of the stream bed in 4 upland streams (Acheron River, Taggerty/Steavenson ri vers, Connelly Creek, and Little River) in southeastern Australia. We measured 25 rocks at each of 32 sites overall, with sites distributed in a nested design: sites were paired in 2nd, 3rd, upper 4th, and lowe r 4th orders on each river. For each rock, we determined whether it wa s wedged into place by surrounding rocks, estimated percentage burial in fine sediments, and measured rock size and ambient water velocity a nd depth. Nested analyses of variance indicated that F, and its correl ates varied most between rivers and greatly between rocks within indiv idual sites; the spatial scales of stream order and site contributed l ittle explanatory power. Hierarchical, log-linear modelling showed tha t both rock size and bed packing varied systematically between rivers, with the Little and Taggerty/Steavenson rivers having relatively larg e rocks that were often packed into the bed, whereas Connelly Creek an d the Acheron River had many relatively-small rocks lying loosely on t op of the bed. A river-by-river analysis showed that Values of F-c wer e related highly to rock sizes but that the nature of the relationship s differed greatly between packed-in rocks and those lying on top of t he bed and also varied between rivers. The Little and Taggerty/Steaven son rivers were similar to each other but both differed from the Acher on River and Connelly Creek, which differed from each other. Our estim ates of F-c suggest that an oft used approximation, which equates tau( c) directly with rock sizes in mm, and the commonly-used equations fro m which the approximation is derived, are likely to produce poor estim ates of tau(c); these poor estimates would cause equally poor estimate s of likely disturbance frequencies. The application by ecologists of reach-level hydraulic equations to estimate shear stresses and the siz es and numbers of rocks moved by floods and spates could be flawed by a focus on inappropriate spatial scales. Our data suggest that Variati on in likely disturbance frequencies between rocks within individual s ites might be of a similar magnitude to variability between different rivers. We argue that spatial variation in stream systems need not be organized in the top-to-bottom hierarchical models that have been rece ntly promoted for rivers.