As any model of real-world phenomena, soil erosion models must be tested against empirical evidence to have their performance evaluated. This is critical to develop knowledge and confidence in model predictions. However, evaluating soil erosion models is complicated due to the uncertainties involved in the estimation of model parameters and measurements of system responses. Here, we undertake a term co-occurrence analysis to investigate how model evaluation is approached in soil erosion research. The analysis illustrates how model testing is often neglected, and how model evaluation topics are segregated from current research interests. We perform a meta-analysis of model performance to understand the mechanisms that influence model predictive accuracy. Results indicate that different models do not systematically outperform each other, and that calibration seems to be the main mechanism of model improvement. We review how soil erosion models have been evaluated at different temporal and spatial scales, focusing on the methods, assumptions, and data used for model testing. We discuss the implications of uncertainty and equifinality in soil erosion models, and implement a case study of uncertainty assessment that enables models to be tested as hypotheses. A comment on the way forward for the evaluation of erosion models is presented, discussing philosophical aspects of hypothesis testing in environmental modelling. We refute the notion that soil erosion models can be validated, and emphasize the necessity of defining fit-for-purpose tests, based on multiple sources of data, that allow for a broad investigation of model usefulness and consistency.

No abstract

Purpose Soils are important regulators of Critical Zone processes that influence the development of geochemical signals used for sediment fingerprinting. In this study, pedological knowledge of tropical soils was incorporated into sediment source stratification and tracer selection in a large Brazilian catchment. Materials and methods In the Ingaí River basin (~1200 km 2), Brazil, three source end-members were defined according to the interpretation of soil and geological maps: the upper, mid, and lower catchment. A tributary sampling design was employed, and sediment geochemistry of three different size fractions was analyzed (2-0.2 mm; 0.2-0.062 mm, and < 0.062 mm). A commonly used statistical methodology to element selection was compared to a knowledge-based approach. The mass balance un-mixing models were solved by a Monte Carlo simulation. Modeled source contributions were evaluated against a set of artificial mixtures with known source proportions. Results and discussion For the coarse fraction (2-0.2 mm), both approaches to element selection yielded high errors compared to the artificial mixtures (23.8% and 17.8% for the statistical and the knowledge-based approach, respectively). The knowledgebased approach provided the lowest errors for the intermediate (0.2-0.062 mm) (10.9%) and fine (< 0.062 mm) (11.8%) fractions. Model predictions for catchment outlet target samples were highly uncertain for the coarse and intermediate fractions. This is likely the result of the spatial scale of the source stratification not being able to represent sediment dynamics for these fractions. Both approaches to element selection show that most of the fine sediments (median > 90%) reaching the catchment outlet were derived from Ustorthents in the lower catchment. Conclusions The different element selection methods and the artificial mixtures provide multiple lines of evidence for evaluating the fingerprint approaches. Our findings highlight the importance of considering pedogenetic processes in source stratification, and demonstrate that different sampling strategies might be necessary to model specific sediment fractions.