Essays on land use modeling and water scarcity

Abstract

Agriculture represents a major use of land across the globe. Half of the world’s habitable land is used for agriculture. In addition, water is an essential factor for life on the planet. Besides being key to human survival, water for irrigation is considered the primary factor for raising the productivity of lands around the world. About 20% of the world’s croplands are irrigated, but they produce around 40% of the global crop supply. Thus, farmers’ decisions with regards to water use in the face of water scarcity is an important topic with economic and environmental implications. Agricultural land decisions are affected by economic factors (e.g., output and input prices) and environmental factors (e.g., climate, soil characteristics, water quality, water availability, ecology). Government policies can have direct and indirect impacts on farmers’ cropping patterns and land use decisions. Water scarcity is a significant constraint on l and use and crop production in the Great Plains. The primary source of irrigation water in Kansas is the High Plains aquifer (HPA), which underlies about 175,000 square miles, in parts of eight states. The HPA has been extensively used for agricultural irrigation in Kansas began in the 1930s and 1940s. Prolonged pumping from this aquifer has led to declining water levels, especially in the southwestern part of the aquifer. Decreases in aquifer levels between 2000 and 2017 ranged from 5 feet to more than 80 feet in southwestern Kansas. The overall purpose of this dissertation is to examine the impact of agricultural land use in Kansas in the context of water scarcity, biofuel expansion, and government policy. The study area is the HPA in Kansas using field-level and county-level data set. Results from these studies provide important contributions to estimation of acreage response and land-use change to water scarcity; the impact of ethanol market expansion on local irrigation decisions; and more robust estimation using a more flexible land use modeling based on land use budgeting. Results of the study are of interest to policy makers, watershed managers, agricultural producers and other Kansas stakeholders interested in land-use and irrigation decisions studies. The first essay is an attempt to examine the effects of water scarcity on crop acreage shares using a multinomial logit (MNL) statistical framework to model land use of major irrigated crops (alfalfa, corn, sorghum, soybeans, and wheat) in the Kansas HPA. We estimate county level and regional level acreage share elasticities with respect to a number of water scarcity variables. Generally, most of the elasticity’s signs are consistent between the county- and regional-level. However, regional-level elasticities are slightly higher in magnitude than elasticities in other studies that do not focus specifically on irrigation. Our results show that water deficit, depth to water, natural gas prices, and quantity of water authorized for extraction have an effect on irrigated crop acreage share at the county- and regional-level. The second essay proposes to demonstrate that a more generalized version of the MNL framework, the nested multinomial logistic (NMNL) statistical framework for modeling land use can provide a flexible model to address a multi crop and multi land use response for assessment of the impacts on land use from economic and non-economic factors. First, we show that NMNL acreage share models can be derived from well-defined profit maximizing behavior under specific assumptions. Second, we show that the NMNL acreage share model can provide a multi-crop and multi-land use econometric modeling framework that can be used to estimate unconditional, conditional and group acreage share elasticities that allow for the assessment of direct and indirect effects of explanatory factors on land use decisions and allocation. The model is illustrated using an empirical example of farmer land use allocation between irrigated cropland, non-irrigated cropland, and non-cropland categories. The third essay aims to examine local irrigation decisions of agricultural producers in the Kansas portion of the HPA in response to ethanol market expansion. To identify the effects of ethanol expansion on local irrigation decisions, we examine field-level data on irrigation water use, total irrigated acreage, and irrigated crop decisions for the years 2000-2018. Specifically, we measure the response of three irrigation decisions: (i) irrigated acreage (extensive margin), (ii) irrigation per acre (intensive margin), and (iii) total water use (total effect) to the location and production capacity of an ethanol plant. We adopt proximity of 25KM and 50KM neighborhood of field from ethanol plant as treatment group to capture the impact of ethanol plant location and capacity on irrigation behavior of fields. We estimate that ethanol plants increase total water use by 12.36 acre-ft and 9.12 acre-ft per field when a new plant is built within 25KM and 50KM neighborhood distance, respectively. In addition, we estimate that a 1 million gallon per year increase in ethanol capacity leads to about a 0.26 and 0.19 acre-ft increase in total water use per field for 25KM and 50KM neighborhood, respectively. Approximately 13% of total response in total water use is due to increases in irrigated acreage. Furthermore, we estimate that a 1 million gallon per year increase in ethanol capacity increases irrigated corn by 0.07 and 0.09 acre per field for 25KM and 50KM neighborhood, respectively.