Search for: All records

Critical loads (CLs) are frequently used to quantify terrestrial ecosystem impacts from nitrogen (N) deposition using ecological responses such as the growth and mortality of tree species. Typically, CLs are reported as a single value, with uncertainty, for an indicator across a species' entire range. Mediating factors such as climate and soil conditions can influence species' sensitivity to N, but the magnitudes of these effects are rarely calculated explicitly. Here, we quantify the spatial variability and estimation error in N CLs for the growth and survival of 10 different tree species while accounting for key environmental factors that mediate species sensitivity to N (e.g., soil characteristics). We used a bootstrapped machine learning approach to determine the level of N deposition at which a 1% decrease occurs in growth rate or survival probability at forest plot locations across the United States. We found minimal differences (<5 kg N ha⁻¹ year⁻¹) when comparing a single species' CLs across climatic regimes but found considerable variability in species' local N CLs (>8.5 kg N ha⁻¹ year⁻¹) within these regimes. We also evaluated the most important factors for predicting tree growth rates and mortality and found that climate, competition, and air pollution generally have the greatest influence on growth rates and survival probability. Lastly, we developed a new probability of exceedance metric for each species and found high likelihoods of exceedance across large portions (46%) of some species' ranges. Our analysis demonstrates that machine learning approaches provide a unique capability to: (1) quantify mediating factor influences on N sensitivity of trees, (2) estimate the error in local N CL estimates, and (3) generate localized N CLs with probabilities of exceedance for tree species. 

The gap between chronological age (CA) and biological brain age, as estimated from magnetic resonance images (MRIs), reflects how individual patterns of neuroanatomic aging deviate from their typical trajectories. MRI-derived brain age (BA) estimates are often obtained using deep learning models that may perform relatively poorly on new data or that lack neuroanatomic interpretability. This study introduces a convolutional neural network (CNN) to estimate BA after training on the MRIs of 4,681 cognitively normal (CN) participants and testing on 1,170 CN participants from an independent sample. BA estimation errors are notably lower than those of previous studies. At both individual and cohort levels, the CNN provides detailed anatomic maps of brain aging patterns that reveal sex dimorphisms and neurocognitive trajectories in adults with mild cognitive impairment (MCI, N  = 351) and Alzheimer’s disease (AD, N  = 359). In individuals with MCI (54% of whom were diagnosed with dementia within 10.9 y from MRI acquisition), BA is significantly better than CA in capturing dementia symptom severity, functional disability, and executive function. Profiles of sex dimorphism and lateralization in brain aging also map onto patterns of neuroanatomic change that reflect cognitive decline. Significant associations between BA and neurocognitive measures suggest that the proposed framework can map, systematically, the relationship between aging-related neuroanatomy changes in CN individuals and in participants with MCI or AD. Early identification of such neuroanatomy changes can help to screen individuals according to their AD risk.