Search

Most useful weather prediction for the public is near the surface. The processes that are most relevant for near-surface weather prediction are also those that are most interactive and exhibit positive feedback or have key role in energy partitioning. Land surface models (LSMs) consider these processes together with surface heterogeneity and forecast water, carbon and energy fluxes, and coupled with an atmospheric model provide boundary and initial conditions. This numerical parametrization of atmospheric boundaries being computationally expensive, statistical surrogate models are increasingly used to accelerated progress in experimental research. We evaluated the efficiency of three surrogate models in speeding up experimental research by simulating land surface processes, which are integral to forecasting water, carbon, and energy fluxes in coupled atmospheric models. Specifically, we compared the performance of a Long-Short Term Memory (LSTM) encoder-decoder network, extreme gradient boosting, and a feed-forward neural network within a physics-informed multi-objective framework. This framework emulates key states of the ECMWF's Integrated Forecasting System (IFS) land surface scheme, ECLand, across continental and global scales. Our findings indicate that while all models on average demonstrate high accuracy over the forecast period, the LSTM network excels in continental long-range predictions when carefully tuned, the XGB scores consistently high across tasks and the MLP provides an excellent implementation-time-accuracy trade-off. The runtime reduction achieved by the emulators in comparison to the full numerical models are significant, offering a faster, yet reliable alternative for conducting numerical experiments on land surfaces.

Despite deep-learning being state-of-the-art for data-driven model predictions, it has not yet found frequent application in ecology. Given the low sample size typical in many environmental research fields, the default choice for the modelling of ecosystems and its functions remain process-based models. The process understanding coded in these models complements the sparse data and neural networks can detect hidden dynamics even in noisy data. Embedding the process model in the neural network adds information to learn from, improving interpretability and predictive performance of the combined model towards the data-only neural networks and the mechanism-only process model. At the example of carbon fluxes in forest ecosystems, we compare different approaches of guiding a neural network towards process model theory. Evaluation of the results under four classical prediction scenarios supports decision-making on the appropriate choice of a process-guided neural network.

Generalised Degrees of Freedom (GDF), as defined by Ye (1998 JASA 93:120-131), represent the sensitivity of model fits to perturbations of the data. As such they can be computed for any statistical model, making it possible, in principle, to derive the number of parameters in machine-learning approaches. Defined originally for normally distributed data only, we here investigate the potential of this approach for Bernoulli-data. GDF-values for models of simulated and real data are compared to model complexity-estimates from cross-validation. Similarly, we computed GDF-based AICc for randomForest, neural networks and boosted regression trees and demonstrated its similarity to cross-validation. GDF-estimates for binary data were unstable and inconsistently sensitive to the number of data points perturbed simultaneously, while at the same time being extremely computer-intensive in their calculation. Repeated 10-fold cross-validation was more robust, based on fewer assumptions and faster to compute. Our findings suggest that the GDF-approach does not readily transfer to Bernoulli data and a wider range of regression approaches., Comment: accompanying R-code on github

Estimating population models from uncertain observations is an important problem in ecology. Perretti et al. observed that standard Bayesian state-space solutions to this problem may provide biased parameter estimates when the underlying dynamics are chaotic. Consequently, forecasts based on these estimates showed poor predictive accuracy compared to simple "model-free" methods, which lead Perretti et al. to conclude that "Model-free forecasting outperforms the correct mechanistic model for simulated and experimental data". However, a simple modification of the statistical methods also suffices to remove the bias and reverse their results., Comment: Letter submitted to PNAS, with additional supplementary information. R code included in the latex source

Ecological networks are often composed of different sub-communities (often referred to as modules). Identifying such modules has the potential to develop a better understanding of the assembly of ecological communities and to investigate functional overlap or specialisation. The most informative form of networks are quantitative or weighted networks. Here we introduce an algorithm to identify modules in quantitative bipartite (or two-mode) networks. It is based on the hierarchical random graphs concept of Clauset et al. (2008 Nature 453: 98-101) and is extended to include quantitative information and adapted to work with bipartite graphs. We define the algorithm, which we call QuaBiMo, sketch its performance on simulated data and illustrate its potential usefulness with a case study., Comment: 19 pages, 10 figures, 4 tables