Search

A data-driven method is presented for improving sampling designs from times series (1D approach) or spatial arrays (2D approach) of digital information. We present the autocorrelated conditioned Latin Hypercube Sampling (acLHS). This method combines a conditioned Latin Hypercube (cLHS) to obtain a representative sample of the joint probability distribution function and an autocorrelation model to reproduce the spatial or temporal dependency function (i.e., temporal or spatial variability). The acLHS method was tested with two case studies using data of soil CO 2 efflux (i.e., the CO 2 flux from soils to the atmosphere known as soil respiration) that are useful for carbon cycle science. First, we used data representing a time series (1D approach), and then spatial data (2D approach) across the conterminous United States (CONUS). Results show that acLHS was more efficient than other sampling methods (i.e., fixed sampling, cLHS) as it better reproduced the joint probability distribution and the temporal or spatial variability of the variable of interest. Finally, we use a Bernstein copula-based stochastic co-simulation method (BCSCS) and demonstrated that the acLHS reduces modeling prediction errors compared with other methods. The acLHS is a flexible method that can be applied to any variable of interest as a time series (1D approach) or as a spatial format (2D approach). • A geostatistical sampling method to improve representativeness of samples. • Implementation works for time series (1D) and spatial information (2D). • The method improves accuracy and precision for prediction purposes. [ABSTRACT FROM AUTHOR]

Annual soil moisture estimates are useful to characterize trends in the climate system, in the capacity of soils to retain water and for predicting land and atmosphere interactions. The main source of soil moisture spatial information across large areas (e.g., continents) is satellite-based microwave remote sensing. However, satellite soil moisture datasets have coarse spatial resolution (e.g., 25–50 km grids); and large areas from regional-to-global scales have spatial information gaps. We provide an alternative approach to predict soil moisture spatial patterns (and associated uncertainty) with higher spatial resolution across areas where no information is otherwise available. This approach relies on geomorphometry derived terrain parameters and machine learning models to improve the statistical accuracy and the spatial resolution (from 27km to 1km grids) of satellite soil moisture information across the conterminous United States on an annual basis (1991–2016). We derived 15 primary and secondary terrain parameters from a digital elevation model. We trained a machine learning algorithm (i.e., kernel weighted nearest neighbors) for each year. Terrain parameters were used as predictors and annual satellite soil moisture estimates were used to train the models. The explained variance for all models-years was >70% (10-fold cross-validation). The 1km soil moisture grids (compared to the original satellite soil moisture estimates) had higher correlations (improving from r2 = 0.1 to r2 = 0.46) and lower bias (improving from 0.062 to 0.057 m3/m3) with field soil moisture observations from the North American Soil Moisture Database (n = 668 locations with available data between 1991–2013; 0-5cm depth). We conclude that the fusion of geomorphometry methods and satellite soil moisture estimates is useful to increase the spatial resolution and accuracy of satellite-derived soil moisture. This approach can be applied to other satellite-derived soil moisture estimates and regions across the world. [ABSTRACT FROM AUTHOR]

Environmental observatory networks (EONs) are coordinated efforts to provide knowledge that ultimately delivers transformational ecological science from regional to global scales. We used ecosystem functional types (EFTs), a time-varying land surface classification, as an alternative way to characterize ecosystem functional heterogeneity based on carbon uptake dynamics. We assessed the representativeness of the eddy-covariance sites of AmeriFlux and NEON, and their combined core sites (i.e., sites with long-term support) across the conterminous United States (CONUS) based on: a) the number of different EFT categories (EFT mode ) represented by each network, b) representativeness of the EFT inter-annual variability (EFT int ; number of unique EFTs within each pixel during years 2001–2014), and c) the spatial representation of EFT mode and EFT int based on a maximum entropy approach (i.e., spatial functional heterogeneity). AmeriFlux represents 50% of all possible EFT categories, includes most of EFT int values (9 out of 14), and represents 55% of the spatial functional heterogeneity across CONUS. NEON represents 23% of all possible EFT categories, 7 out of 14 possible EFT int values, and 23% of the spatial functional heterogeneity across CONUS. The combined effort of AmeriFlux and NEON core sites represents 33% of all possible EFT categories, 7 out of 14 possible EFT int values, and 46% of the spatial functional heterogeneity across CONUS. We used the NEON ecoclimatic domains to summarize our results within a geographical context. The least represented NEON ecoclimatic domains were Desert Southwest, Southern Rockies and Colorado Plateau, Great Basin, Northern Plains, and Central Plains. Our results provide insights about the potential of AmeriFlux to address questions regarding decadal and inter-annual variability of ecosystem functional heterogeneity across CONUS. [ABSTRACT FROM AUTHOR]