American Indian reservations have low incomes and high rates of poverty relative to adjacent communities, and the income gap appears to be even larger for Indian farmers. We examine the extent to which a lack of access to capital might explain these differences using irrigation systems as a proxy for on‐farm investment around the Uintah‐Ouray Indian Reservation in eastern Utah. Uintah land is held in trust by the US government, and farmers on this land face significant barriers to acquiring capital to invest in irrigation equipment and infrastructure. We use the boundaries from a 1905 land allotment as a natural experiment, employing both sharp and fuzzy regression discontinuity designs to explore whether agricultural land use, irrigation levels, irrigation investment, and crop choice differ across the boundary. The original allocation provided similar land in the immediate neighborhood around its borders, and our results suggest that today tribal trust land is farmed and irrigated at rates similar to adjacent land. However, conditional on being irrigated, tribal trust land is around thirty‐two percentage points less likely to utilize capital‐intensive sprinkler irrigation, and up to ten percentage points less likely to grow high‐value crops. Trust ownership, which is characterized by cumbersome bureaucratic processes, limits on agricultural lease flexibility, and the inability to use land as collateral to acquire loans, is a likely explanation for the observed differences.
Seismic impedance inversion can be performed with a semi-supervised learning algorithm, which only needs a few logs as labels and is less likely to get overfitted. However, classical semi-supervised learning algorithm usually leads to artifacts on the predicted impedance image. In this artical, we improve the semi-supervised learning from two aspects. First, by replacing 1-d convolutional neural network (CNN) layers in deep learning structure with 2-d CNN layers and 2-d maxpooling layers, the prediction accuracy is improved. Second, prediction uncertainty can also be estimated by embedding the network into a Bayesian inference framework. Local reparameterization trick is used during forward propagation of the network to reduce sampling cost. Tests with Marmousi2 model and SEAM model validate the feasibility of the proposed strategy.
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.