The prespecification of the network is one of the biggest hurdles for applied researchers in undertaking spatial analysis. In this letter, we demonstrate two results. First, we derive bounds for the bias in nonspatial models with omitted spatially-lagged predictors or outcomes. These bias expressions can be obtained without prior knowledge of the network, and are more informative than familiar omitted variable bias formulas. Second, we derive bounds for the bias in spatial econometric models with nondifferential error in the specification of the weights matrix. Under these conditions, we demonstrate that an omitted spatial input is the limit condition of including a misspecificed spatial weights matrix. Simulated experiments further demonstrate that spatial models with a misspecified weights matrix weakly dominate nonspatial models. Our results imply that, where cross-sectional dependence is presumed, researchers should pursue spatial analysis even with limited information on network ties.

This study evaluates the effects of vegetative soil conservation practices (afforestation and/or bamboo planting) on farm profit and its components, revenue and variable cost. Since farmers self-select themselves as adopters of conservation measures, there could be a problem of selection bias in evaluating their soil conservation practices. We address the selection bias by using propensity score matching. We also check if there exists spatial spillover in adoption of vegetative conservation measures and how it affects matching. We use primary survey data from the Darjeeling district of the Eastern Himalayan region for the year 2013. Our results suggest strong spatial correlation. We find that the propensity score estimated from the spatial model provides better matches than the non-spatial model. While the results show that vegetative soil conservation can lead to significant gains in revenue, it also increases costs so that no significant gains in profit accrue to farmers.

Using remotely sensed land-cover data in 1994 and 2014, and cross-sectional survey data in 2014, this study examines the association between land use and cover change and agricultural productivity in northern Ghana. We document a significant expansion of crop land and settlements (productive use) at the expense of natural vegetation cover. Land areas converted from natural cover to productive use have higher maize yield (0.17 tons per hectare) and harvest value (1,021 Ghanaian Cedi) compared with those converted from bare soil to productive cover. Moreover, areas that were covered by shrubs or savannah in 1994 were more productive in 2014 relative to bare soils in 1994. Although our data do not allow us to establish causality, the evidence suggests the importance of past land-cover conditions in affecting current agricultural performance, especially in resource-stricken settings where conservation and restoration practices are not as common.

Controlling for spatial heterogeneity and spatial dependence in farmland valuation models has gained substantial attention in recent literature. This paper proposes to derive the spatial structure of farmland values endogenously and semiparametrically based on the spatial competition theory. The paper assembles panel data of Pennsylvania county level farmland values between 1982 and 2012. A spatial autoregressive panel data model with spatial weights matrix endogenously incorporated is estimated. Out of sample predictions and non-nested statistical tests for model selection suggest that the fit and the predictability of hedonic farmland valuation models can be greatly improved.

Cross-border commercial activity raises issues in federations where multiple jurisdictions can claim the right to tax the same income. In the United States, this coordination problem is resolved by splitting the tax base according to the geographic distribution of firms’ sales, capital and labour. The weight of each factor is determined on a state-by-state basis, which opens room for competitive legislative behaviour. In this complex issue area, however, policymakers must invest lot of resources to monitor competitors, evaluate policy alternatives and shepherd tax reform through the legislative process. This implies that highly professional legislatures should be more responsive to the policies of nearby states. We consider data on most American states over the period from 1986 to 2013 and find strong evidence of conditional spatial dependence. Our findings suggest that policy diffusion may often be moderated by institutional and political factors.

Predictions of weed seedling populations from seedbank data should characterize the spatial distribution as well as the composition and abundance of weeds. The spatial distribution of seedbank and seedling populations of common lambsquarters and annual grasses (giant foxtail, large crabgrass, and fall panicum) were described in moldboard plow and no-tillage soybean fields from 1990 to 1993. Spearman rank correlations between seedbank and seedling densities were significant for common lambsquarters in both tillages and all years, but for annual grasses correlations were significant only in no-tillage. Semivariograms showed spatial autocorrelation in seedbank and seedling populations of common lambsquarters in all years in no-till, but less often in the moldboard plow field. Annual grass seed and seedling populations were autocorrelated in the no-till field every year except 1993, and in the moldboard plow field in 1992 and 1993 only. Cross-semivariograms showed spatial continuity between seedbank and seedling population densities in 3 of 4 yr in no-till for common lambsquarters, and in all years of no-till and 1 yr of moldboard plow for annual grasses. Grey-scale field maps of common lambsquarters seedbanks corresponded visually to maps of seedling populations and could have been used to target control efforts, but visual correspondence between annual grass seedbank and seedling maps was poor. Seedbank and seedling mapping may be useful for site-specific management, but additional information is needed to understand the variation in the relationships between these two populations over time and space.

Geostatistical methods were used to describe and map nonrandom distribution and variation (standard deviation) of shoot density and root growth across a well-established patch of Canada thistle, a perennial weed. Semivariogram functions and kriging, an interpolation method, were used to prepare isoarithmic contour maps and associated error maps. Maps consisted of interpolated contours of uniform weed density and other measured or calculated regionalized variables between measured X-Y control points, as well as maps of error (standard deviation) associated with contour estimation. Mapped regions of greatest shoot density across a patch not only had the greatest underlying root biomass and, often, greatest density of adventitious root buds, but also had more deeply growing root biomass.

In non-life insurance, territory-based risk classification is useful for various insurance operations including marketing, underwriting, ratemaking, etc. This paper proposes a spatially dependent frequency-severity modeling framework to produce territorial risk scores. The framework applies to the aggregated insurance claims where the frequency and severity components examine the occurrence rate and average size of insurance claims in each geographic unit, respectively. We employ the bivariate conditional autoregressive models to accommodate the spatial dependency in the frequency and severity components, as well as the cross-sectional association between the two components. Using a town-level claims data of automobile insurance in Massachusetts, we demonstrate applications of the model output–territorial risk scores–in ratemaking and market segmentation.

Comparing distributions among fields, species, and management practices will help us understand the spatial dynamics of weed seed banks, but analyzing observational data requires nontraditional statistical methods. We used cluster analysis and classification and regression tree analysis (CART) to investigate factors that influence spatial distributions of seed banks. CART is a method for developing predictive models, but it is also used to explain variation in a response variable from a set of possible explanatory variables. With cluster analysis, we identified patterns of variation with direction of the distance over which seed bank density was correlated (range of spatial dependence) with single-species seed banks in corn. Then we predicted patterns of the seed banks with CART using field and species characteristics and seed bank density as explanatory variables. Patterns differed by magnitude of variation in the range of spatial dependence (strength of anisotropy) and direction of the maximum range. Density and type of irrigation explained the most variation in pattern. Long ranges were associated with large seed banks and stronger anisotropy with furrow than center pivot irrigation. Pattern was also explained by seed size and longevity, characteristics for natural dispersal, species, soil texture, and whether the weed was a grass or broadleaf. Significance of these factors depended on density or type of irrigation, and some patterns were predicted for more than one combination of factors. Dispersal was identified as a primary process of spatial dynamics and pattern varied for seed spread by tillage, wind, or natural dispersal. However, demographic characteristics and density were more important in this research than in previous research. Impact of these factors may have been clearer because interactions were modeled. Lack of data will be the greatest obstacle to using comparative studies and CART to understand the spatial dynamics of weed seed banks.

The knowledge of weed distribution in a field is a key factor to manage weeds effectively. The feasibility of using weed distribution maps for site-specific weed control will largely depend on the stability of the spatial distribution of the populations. Seed banks are the most reliable way of telling the area's weediness, but the effect of regular herbicide applications on its stability is largely unknown. A field experiment was conducted during 3 yr in a winter wheat field under herbicide treatments with the aim of studying the seed bank's spatial distribution of prostrate knotweed and corn poppy and the spatiotemporal stability of their populations. Soil samples were taken each year on the same locations, and seed abundance was measured by germination in greenhouse. Both species accounted for more than 10% of the broad-leaved weed seed bank and they were selected for further analysis. Prostrate knotweed seed-bank density decreased 76% and corn poppy 88% in 3 yr. Spatial distribution was described by spherical isotropic semivariograms. Distance of spatial dependence (range) of prostrate knotweed and corn poppy decreased 33 and 11% respectively, and the spatial variability (sill) decreased 96 and 99%. Yearly spatial seed distribution was compared for each species and no temporal stability was observed over a 3-yr period. The lack of stability was attributed to the important decrease of seed density over time and the increase in the short-range variability (nugget). However, for prostrate knotweed, the location of minima and maxima were roughly the same between years, allowing farmers to extend the period of use of the weed distribution maps. Although spatial distribution of seed banks can be affected by processes that promote fast changes in the densities of weed populations, this fact does not mean that weed distribution maps could not be used in consecutive seasons.

Growers need affordable methods to sample weed populations to reduce herbicide use with site-specific weed management. Sampling programs and methods of developing sampling programs for integrated pest management are not sufficient for site-specific weed management because more and different information is needed to make treatment maps than simply estimate average pest density. Sampling plans for site-specific weed management must provide information to map the weeds in the field but should be developed for the objective of prescribing spatially variable management. Weed scientists will be most successful at designing plans for site-specific weed management if they focus on this objective throughout the process of designing a sampling plan. They must also learn more about the spatial distribution and dynamics of weed populations and use that knowledge to identify cost-effective plans, recommend methods to make maps as well as collect data, and find ways to evaluate maps that reflect management to be prescribed from the map. Foremost, sampling must be thought of as an ongoing process over time that uses many types of information rather than a single event of collecting one type of information. Specifically, scientists will need to identify common characteristics rather than just differences of the spatial distribution of weeds among fields and species, recognize that map accuracy may be a poor indicator of the value of a sampling plan, and develop methods to use growers' knowledge of the distribution of weeds and past spatially variable management within a field for both making a map and recommending a sampling plan. The value of proposed methods for sampling and mapping must also be demonstrated or adoption of site-specific weed management might be limited to growers who enjoy using sophisticated technology.

Spatial analysis can be used to relate the patterns of tree species to their regeneration syndromes – pioneer to late-successional – and is a first step in refining hypotheses about the species traits and biotic and abiotic factors that give rise to forest community dynamics. This study examines the spatial pattern of the most abundant trees in three 0.45-ha plots in species-poor lowland rain forests on oceanic islands in Tonga, Western Polynesia, that experience frequent natural disturbance and have a 3000-y history of shifting cultivation. We contrast secondary vs. remnant late-successional forest, with particular attention paid to the spatial dispersion and clustering of tree species, and the presence of spatial dependence in the density of seedlings and saplings. Shade-tolerant species were not strongly clustered at any scale. They did not appear to be dispersal limited, in late successional forest, and only some showed patterns consistent with density-dependent mortality (more clumped when small). Shade-tolerant species were more clumped in secondary forest, and may be dispersal-limited there because vertebrate dispersers prefer primary forest. Shade-intolerant species were clumped in gaps in late-successional forest, but some were also clumped in secondary forest, indicating that they too may be dispersal limited during secondary succession. We also compared the species composition of seedlings and saplings in the centre of plots with trees in the surrounding area and inferred that active dispersal (by vertebrate frugivores) contributed as much as 50% to site species richness.