Daisuke Murakami

Random effects specifications in eigenvector spatial filtering: a simulation study

Eigenvector spatial filtering (ESF) is becoming a popular way to address spatial dependence. Recently, a random effects specification of ESF (RE-ESF) is receiving considerable attention because of its usefulness for spatial dependence analysis considering spatial confounding. The objective of this study was to analyze theoretical properties of RE-ESF and extend it to overcome some of its disadvantages. We first compare the properties of RE-ESF and ESF with geostatistical and spatial econometric models. There, we suggest two major disadvantages of RE-ESF: it is specific to its selected spatial connectivity structure, and while the current form of RE-ESF eliminates the spatial dependence component confounding with explanatory variables to stabilize the parameter estimation, the elimination can yield biased estimates. RE-ESF is extended to cope with these two problems. A computationally efficient residual maximum likelihood estimation is developed for the extended model. Effectiveness of the extended RE-ESF is examined by a comparative Monte Carlo simulation. The main findings of this simulation are as follows: Our extension successfully reduces errors in parameter estimates; in many cases, parameter estimates of our RE-ESF are more accurate than other ESF models; the elimination of the spatial component confounding with explanatory variables results in biased parameter estimates; efficiency of an accuracy maximization-based conventional ESF is comparable to RE-ESF in many cases.

Application of Lasso to the Eigenvector Selection Problem in Eigenvector Based Spatial Filtering

Eigenvector based spatial filtering is one of the well-used approaches to model spatial autocorrelation among the observations or errors in a regression model. In this approach, subset of eigenvectors extracted from a modified spatial weight matrix is added to the model as explanatory variables. The subset is typically specified via the forward stepwise model selection procedure, but it is disappointingly slow when the number of observations n takes a large number. Hence as a complement or alternative, the present paper proposes the use of the LASSO (L1-penalized regression) to select the eigenvectors. The LASSO model selection procedure is applied to the well-known Boston housing dataset and simulation dataset, and its performance is compared with that of the stepwise procedure. The obtained results suggest that the LASSO is fairly fast compared the stepwise procedure, and can select eigenvectors effectively even if dataset is relatively large (n = 10000), to which the forward stepwise procedure is uneasy to apply.

Participatory Sensing Data Tweets for Micro-Urban Real-Time Resiliency Monitoring and Risk Management

Real-time urban climate monitoring provides useful information that can be utilized to help urban management personnel to monitor and adapt their precautionary measures to extreme events, including urban heatwaves. Fortunately, recently created social media platforms, such as Twitter, furnish real-time and high-resolution spatial information that may be useful for climate condition estimation. The objective of this paper was to utilize geotagged tweets (participatory sensing data) for urban temperature analysis. We first detected tweets related to heat (heat-tweets). Then, we examined the relationships between monitored temperatures and heat-tweets through a statistical model framework based on copula modeling methods. We demonstrate that there are strong relationships between heat-tweets and temperatures recorded at an intra-urban scale, which are revealed by our analysis of Tokyo city and its suburbs. Subsequently, we investigated the application of heat-tweets for informing spatiotemporal Gaussian process interpolation of temperatures as an application example of heat-tweets. We utilized a combination of spatially sparse weather monitoring sensor data, which comprise infrequently available moderate resolution imaging spectroradiometer remote sensing data and spatially and temporally dense lower quality geotagged Twitter data. A spatial best linear unbiased estimation technique was applied. The results suggest that tweets provide the useful auxiliary information for urban climate assessment.

A first step towards resilient graph partitioning for electrical grids

We study a graph partitioning problem for electrical grids such that a given grid is partitioned into multiple ones that are self-contained concerning electricity balance. Our goal is to find a resilient partition against time-changing power demand and supply over the year. In this paper, we investigate two graph partitioning algorithms applying them to a synthesized dataset based on realistic assumptions about Yokohama, Japan. Our initial results show that a simple algorithm, which only considers horizontal or vertical partitions, possibly produces more resilient partitions than a more general algorithm whose partitions divide a graph into subgraphs of any topology.

Area-to-point parameter estimation with geographically weighted regression

The modifiable areal unit problem (MAUP) is a problem by which aggregated units of data influence the results of spatial data analysis. Standard GWR, which ignores aggregation mechanisms, cannot be considered to serve as an efficient countermeasure of MAUP. Accordingly, this study proposes a type of GWR with aggregation mechanisms, termed area-to-point (ATP) GWR herein. ATP GWR, which is closely related to geostatistical approaches, estimates the disaggregate-level local trend parameters by using aggregated variables. We examine the effectiveness of ATP GWR for mitigating MAUP through a simulation study and an empirical study. The simulation study indicates that the method proposed herein is robust to the MAUP when the spatial scales of aggregation are not too global compared with the scale of the underlying spatial variations. The empirical studies demonstrate that the method provides intuitively consistent estimates.

Practical spatial statisics for areal interpolation

Differences in spatial units among spatial data often complicate analyses. Spatial unit conversion, called areal interpolation, is often applied to address this problem. Of the many proposed areal interpolation methods, few consider spatial autocorrelation, which is the general property of spatial data. In this paper an areal interpolation method is constructed by combining a spatial process model, a primal model in spatial statistics, and the linear-regression-based areal interpolation method. The primal advantages of our methods are twofold: It considers both spatial autocorrelation and the volume-preserving property; it is more practical than other spatial-statistics-based areal interpolation methods. A case study on the areal interpolation of the density of employee numbers is provided to check the properties of our method. This case study shows that our method succeeds in improving predictive accuracy. Furthermore, the areal interpolation result indicates that our method, which provides a smooth interpolation map, is appropriate to model the underlying process of spatially aggregated data. These results indicate that the consideration of spatial autocorrelation is important for areal interpolation.

New Spatial Econometrics–Based Areal Interpolation Method:

Spatial data are often aggregated into spatial units and differences between spatial units can complicate the analysis of the data. One solution to this problem is spatial unit conversion, also called areal interpolation. Of the many areal interpolation methods proposed thus far, few method are based on spatial econometrics: a subset of econometrics which is concerned with the role of spatial autocorrelation (a general property of spatial data that implies that data in nearby locations are similar) in the regional economic model response. In this article, an areal interpolation method that considers both the spatial autocorrelation and the pycnophylactic property (a most basic premise of areal interpolation that the sum of the data given in a specific area must be constant) is proposed by combining a spatial econometric model and a linear regression-based areal interpolation method. Parameters of the proposed method are estimated using the expectation-maximization algorithm. The performance of the proposed method was examined through empirical analysis using real data and ratios on aging populations. The results indicate the importance of considering both the pycnophylactic property and the spatial autocorrelation in areal interpolation. The results also show the applicability of spatial econometrics to areal interpolation problems.

Estimating water–food–ecosystem trade-offs for the global negative emission scenario (IPCC-RCP2.6)

Negative emission technologies such as bioenergy with carbon capture and storage (BECCS) are regarded as an option to achieve the climatic target of the Paris Agreement. However, our understanding of the realistic sustainable feasibility of the global lands for BECCS remains uncertain. In this study, we assess the impact of BECCS deployment scenarios on the land systems including land use, water resources, and ecosystem services. Specifically, we assess three land-use scenarios to achieve the total amount of 3.3xa0GtCxa0year−1 (annual negative emission level required for IPCC-RCP 2.6) emission reduction by growing bioenergy crops which requires huge use of global agricultural and forest lands and water. Our study shows that (1) vast conversion of food cropland into rainfed bio-crop cultivation yields a considerable loss of food production that may not be tolerable considering the population increase in the future. (2) When irrigation is applied to bio-crop production, the bioenergy crop productivity is enhanced. This suppresses the necessary area for bio-crop production to half, and saves the land for agricultural productions. However, water consumption is doubled and this may exacerbate global water stress. (3) If conversion of forest land for bioenergy crop cultivation is allowed without protecting the natural forests, large areas of tropical forest could be used for bioenergy crop production. Forest biomass and soil carbon stocks are reduced, implying degradation of the climate regulation and other ecosystem services. These results suggest that without a careful consideration of the land use for bioenergy crop production, a large-scale implementation of BECCS could negatively impact food, water and ecosystem services that are supporting fundamental human sustainability.

Resilient Community Clustering: A Graph Theoretical Approach

Many complex systems can be modeled as a graph consisting of nodes and connecting edges. Such a graph-based model is useful to study the resilience of decentralized systems that handle a system failure by isolating a subsystem with failed components. In this chapter, we study a graph clustering problem for electrical grids where a given grid is partitioned into multiple microgrids that are self-contained in terms of electricity balance. Our goal is to find an optimal partition that minimizes the cost of constructing a set of self-sufficient microgrids. To obtain a better solution accommodating smaller microgrids, we develop an efficient verification algorithm that determines whether microgrids can balance their electricity surplus through electricity exchange among them. Our experimental results with a dataset about Yokohama city in Japan show that our proposed method can effectively reduce the construction cost of decentralized microgrids.

Spatially Explicit Land-Use Modelling for Assessing Climate-Resilient Sustainable Urban Forms

Climate resiliency is a key topic for cities across the world especially after the Paris Agreement. By combing socio-economic situations, carbon emission reduction, disaster risk management, and other factors determine sustainability of cities, we need to understand trade-offs among these factors. In other words, wise urban systems design are required in cities in the world. Especially in the developed countries like Japan which are expected to experience unprecedented pollution decrease, we need to achieve “wise shrink” of cities that are desirable from multiple viewpoints. With such a background, we introduce our spatially-explicit urban land-use model (SULM) as a tool to analyze the trade-offs among climate resilient sustainability. The SULM is applied to an analysis in the Tokyo metropolitan area to analyze the implications of Business-As-usual (BAU), Compact city, and Wise shrink land use scenarios. SULM could be a useful tool for assessing urban resiliency against climate extreme events for eco-urbanism oriented land forms.