Journal Paper Digests

Reading time ~5 minutes

Journal Paper Digests 2022 #15

  • Calibration set optimization and library transfer for soil carbon estimation using soil spectroscopy-A review
  • Figures, landscapes and landsystems: Digital locations, connectivity and communications COMMENT
  • Sub-sampling a large physical soil archive for additional analyses to support spatial mapping; a pre-registered experiment in the Southern Nations, Nationalities, and Peoples Region (SNNPR) of Ethiopia
  • A scalable method for the estimation of spatial disaggregation models

A scalable method for the estimation of spatial disaggregation models

Gaining information about detailed processes using aggregation information is a frequent challenge in research involving geospatial data, with examples in different fields of knowledge such as agronomy, soil science, meteorology, public health, epidemiology, and others. Analyses using aggregated data lead to distorted conclusions since they disregard local patterns, and such a problem has motivated different approaches for reconstructing the information in a finer resolution from the aggregated data. However, most existing methods focus on the particular case where the volume of data does not exceed the amount of memory available for computations, a situation that has become increasingly less frequent with the fast pace of data generation nowadays. In practice, this problem limits either spatial resolution or coverage of applications, thus precluding their use in a more general context. In this paper, we address the problem of disaggregation of spatial data with huge datasets by proposing a scalable method to estimate the parameters of a well-established model. We propose an iterative scheme for model estimation and prove its convergence to a critical point of the likelihood function derived. To test the method, we provide a controlled simulation and a real example for sugarcane production in Brazil. In the simulation, the results indicate a successful reconstruction of 1 million pixels from 90 block areas. In the real example, the results had a compatible match with the agronomic literature, indicating a reasonable prediction of sugarcane production in a 100 m spatial resolution (i.e., approx. 5 x 10(8) pixels) from 5,565 block-areas. Compared to the most similar previous work, scalability allowed us to use a nearly 100 times higher resolution, which corresponds to 10,000 times more pixels. With our methods, we expect to assist researchers from different fields in disaggregating spatial information to larger areas or higher resolutions.

Sub-sampling a large physical soil archive for additional analyses to support spatial mapping; a pre-registered experiment in the Southern Nations, Nationalities, and Peoples Region (SNNPR) of Ethiopia

The value of physical archives of soil material from field sampling activities has been widely recognized. If we want to use archive material for new destructive analyses to support a task, such as spatial mapping, then an efficient sub-sampling strategy is needed, both to manage analytical costs and to conserve the archive material. In this paper we present an approach to this problem when the objective is spatial mapping by ordinary kriging. Our objective was to subsample the physical archive from the Ethiopia Soil Information System (EthioSIS) survey of the Southern Nations, Nationalities and Peoples Region (SNNPR) for spatial mapping of two variables, concentrations of particular fractions of selenium and iodine in the soil, which had not been measured there. We used data from cognate parts of surrounding regions of Ethiopia to estimate variograms of these properties, and then computed prediction error variances for maps in SNNPR based on proposed subsets of the archive of different size, selected to optimize a spatial coverage criterion (with some close sample pairs included). On this basis a subsample was selected. This is a preregistered experiment in that we have proposed criteria for evaluating the success of our approach, and are publishing that in advance of receiving analytical data on the subsampled material from the laboratories where they are being processed. A subsequent short report will publish the outcome. The use of preregistered trials is widely recommended and used in areas of science including public health, and we believe that it is a sound strategy to promote reproducible research in soil science.

Figures, landscapes and landsystems: Digital locations, connectivity and communications COMMENT

Current geomorphological research tends to give preference to landforms rather than landscapes. This commentary brings these notions closer together and incorporates varied views of ‘landscape’ (such as landscapes in art and poetry) as well as showing the difficulties of communicating them to audiences wider than geomorphology. The paper promotes the use of decimal Latitude Longitude (dLL) values to assist better location of samples, analyses and especially landforms, to produce ‘information surfaces’ within landscape and landsystem approaches to geomorphology. This will also help the development of geomorphic visual literacy.

Calibration set optimization and library transfer for soil carbon estimation using soil spectroscopy-A review

Resource-efficient techniques for accurate soil property estimation are necessary to satisfy the increasing demand for soil data to support environmental monitoring, precision agriculture, and spatial modeling. Over the last 30 yr, infrared soil spectroscopy has developed into a rapid, robust, and cost-effective technique for soil carbon analysis. Ongoing global efforts to make soil spectroscopy operational require the development of soil spectral libraries, which are the main source of data for the construction of calibration models. Understanding calibration optimization is important to ensure the efficient use of soil spectral libraries for the accurate estimation of soil carbon. Moreover, spectral library transfer can benefit new data collection, soil monitoring, and modeling efforts. This review presents techniques for optimization of calibration models and library transfer. Selection of calibration set size and subsetting are presented as current calibration optimization techniques. Moreover, spiking is discussed as an effective technique for spectral library transfer. Overall, studies have suggested that an increase in calibration size improves model performance and this continues until an optimal size is reached. Additionally, subsetting can improve model performance if the resulting subsets reduce the variability of spectrally active components. Studies have also suggested that spiking is effective when used in conjunction with subsetting techniques. These findings denote the current applicability and potential of optimization and library transfer techniques for the accurate estimation of soil carbon with soil spectroscopy. Future efforts should focus on refining optimization techniques to further expand the operability of soil spectroscopy for soil carbon estimation.

Journal Paper Digests

Journal Paper Digests 2022 #14 Comparison of bagging, boosting and stacking algorithms for surface soil moisture mapping using optical-t...… Continue reading

Journal Paper Digests

Published on July 17, 2022

Journal Paper Digests

Published on July 10, 2022