For growers facing highly saline groundwater, the recycled water can also serve as a higher quality water source, allowing for the production of more salt-sensitive crops and increasing yields. While recycled municipal water is a promising concept, and programs are starting to be implemented in water-stressed regions, there is little economic research on the impacts of recycled water to growers. This study explores the value of recycled water in two ways: the welfare gains for growers directly receiving water deliveries, through higher-value crop choices and improved yields; and changes in the underlying aquifer quality below the recycled water delivery area, to evaluate indirect impacts of recycled water. To study the direct effects of recycled water on crop choices and welfare, I use a panel mixed logit choice model. The structure of this approach allows for simulations of grower decisions in the absence of a recycled water program and estimates of the willingness to pay for changes in groundwater salinity, while capturing heterogeneity in crop choice decisions. To evaluate the impacts of recycled water on aquifer salinity, I use staggered difference-in-differences and event studies, vertical agriculture given the slow ramp-up of water deliveries over time. I find that growers receiving recycled water are less likely to leave their ground fallow, and are more likely to plant high-value, salt sensitive crops, such as strawberries.

Welfare gains for producers directly receiving water deliveries add up to $16 million annually for 5500acre-ft of delivered water. I also find that recycled water improves the water quality of the underlying aquifer. The effects are strongest directly underneath parcels receiving recycled water, and are driven by years where salinity levels are highest. Neighboring parcels that do not receive water deliveries only see improvements in water quality in high-salinity drought years. Since recycled water also has non-negligible salinity levels, improvements to water quality are negligible or non-existent in years of high rainfall. Altogether, these welfare gains are a conservative estimate for benefits to agricultural producers. I look at immediate benefits to agricultural producers, as they make crop choice decisions in a saline, water-constrained environment. This study does not capture the full effects of the additional water supply in preventing longer-term seawater overdraft. For example, farmland further away from the delivered water zone may be able to avoid severe seawater intrusion in upcoming decades due to the recycled water program, although their current salinity levels may not change. However, the findings of improvements in water quality directly on the coast do serve as promising indicators that future seawater intrusion is being at least partially managed. While beyond the scope of this paper, if hydrologic models were used to estimate individual well-level externalities, a basin-wide model of improvements in water quality may be estimated.

The combination of Pajaro Valley’s high value crops and temperate coastal climate means that they are willing to invest in expensive programs to maintain agricultural production in the region. However, the future of using municipal treated wastewater as an additional water source is expanding rapidly. There are 250 small-scale recycled water programs in California alone, with others in Arizona, Florida, and Texas. Moreover, there is a promising future for recycled water in locations such as in South Korea, where groundwater overdraft has led to seawater intrusion in 41-50% of coastal groundwater sources . Recycled water has a three-fold benefit in our setting and in many other coastal regions, where it provides an additional water supply, improves water quality, and prevents seawater intrusion. However, in the face of future drought and limited freshwater availability, the benefits of an additional water supply may alone be enough to justify the costs of a recycled water facility in arid regions worldwide. Regardless, recycled water may play a major role in climate change mitigation strategies for coastal regions in the future.Insect pollination is important to commercial blueberry production, and farmers usually rent honeybee hives and, when possible, also buy Bombus terrestris L. colonies and place them in the field during the flowering season. Bumblebees are very effective pollinators for this crop compared to other bees . They can work relatively early in the season when most blueberries flower, at cool temperatures unfavorable to honeybee pollination .According to a review by Garibaldi et al. the proximity of natural areas can increase bumblebee density in crop areas since it can provide floral resources and/or undisturbed nesting/overwintering habitat.

Floral resources provided by crop and non-crop areas can also increase bumblebee densities , but the temporal dynamics of flowering crops alters bumblebee densities as well. Previous studies reported a ‘transient dilution effect’ in which bumblebee densities decrease with increasing area of oilseed rape fields during flowering, both in this crop area and in nearby grassland areas . Only after the flowering events of oilseed rape bumblebee densities increased in nearby areas . Based on these observations, it is possible that bumblebee density in blueberry crops is positively influenced by natural areas and negatively by simultaneously flowering crops in the surrounding landscape. Research questions of this study were: Are there any effects of surrounding natural areas and flowering crop areas on wild bumblebee abundance associated with blueberry fields? What is the relevant spatial scale for these effects? Increasing the understanding of landscape effects on important wild pollinators can provide cues on possible management strategies to increase pollination services for blueberry production. Eight blueberry farms in the central valley of the Region La Araucanía, Chile, were studied . Samples were collected from ‘Briggita’ cultivar of high bush blueberry . Distance between farms ranged from 7.4 to 97.8 km . Blueberry fields varied in size from 0.5 to 120 ha . Commercial honeybee colonies were employed in 6 farms while 4 of those also had B. terrestris colonies. None of the surveyed farms applied pesticides during flowering. Weeds were suppressed in and around all of the fields. Sampling of pollinators was performed on two different dates within the flowering season, between October 13th and November 5 th , 2011, between 11:00 to 17:00 on days with favorable weather conditions . All sampled fields had the same plant density . Flower density was assumed to be relatively constant as blueberry variety was the same and plant age was relatively homogeneous . Depending on the field size, 4, 5 or 6 sites per farm were sampled; in one case only 1 site was sampled because of the limited field size . Sampled sites were placed at different locations within the fields including field edges and field centers, in order to capture possible variation of pollinator densities. Each sampling event consisted of separate counts of 4 successive rows in which a person walked through 10 consecutive plants in a row for 5 minutes,vertial farming aeroponics recording all insects that visited flowers. All row counts were summed to produce a site-level estimate of abundance.Sampling was performed when the proportion of open flowers on 10 randomly chosen branches was ≥ 0.2. Bumblebees and honeybees were visually identified to the species level and other pollinators were recorded as other hymenopterans, syrphids, or other. During sampling wild bumblebee workers were not yet active and thus all sampled workers were assumed to be from commercial colonies placed by farmers. At the time of sampling no workers were observed in the area, except for those provided by commercial colonies.Natural forest areas and high-food-resources areas surrounding the blueberry fields were mapped in a radius of 3.50 km from the center of the fields, based on orthorectified aerial photos acquired between 2008 and 2010 depending on the location, and high resolution imagery available from Google Earth . The software ArcGIS 9.3 was used for this purpose. Natural forest areas correspond to unmanaged secondary forests and don’t include exotic pine and eucalyptus plantations. High-food resources areas were mapped using the imagery and identified via visual inspection as food resource if flowering in the mapped area at the time of the surveys. The influence of landscape context on pollinator abundance was analyzed at different spatial scales using circular neighborhoods centered on the central point of sampled sites within the farm with radii of 0.5, 1.0, 1.5, 2.0, 2.5, 3.0, and 3.5 km.

To explore larger landscape effects circular areas of 5 and 8 km radius were used from on a land cover map of 15 m spatial resolution. This map was generated by classification of an ASTER satellite image from 2008 and was used to measure only natural forest areas at these scales. One farm at these scales of analysis was excluded because its buffer overlapped with neighboring farms. The proportion of natural forest and high-food-resources areas varied across farms and both tended to decrease with increasing spatial scale .Analysis of landscape effects focused on the relationship between the proportion of forest area and high-food-resources area with the abundance of wild B. terrestris queens. This group was the most abundant among sampled pollinators.Linear regression models to predict farm-level bumblebee abundance were fitted as a function of the proportion of forest area and high-food-resources area, accounting for unequal number of sites per farm weighting each observation by the number of sites of that farm. Models with two covariates were tested at each scale of analysis considering also all combinations of different spatial scales. In all models loge-loge transformation of the data was used to improve linearity. Correlation between forest and high-food resources areas at each spatial scale was tested using Pearson correlation coefficient. R v.2.15.0 was used for statistical analysis. The possible effect of commercial bumblebee and honeybee colonies was not evaluated. These colonies are placed in the field only during blueberry flowering and competition for food resources is unlikely given the over-abundance of this resource. Honeybees were the most abundant pollinator sampled, followed by naturally occurring B. terrestris queens and syrphids . Few individuals of B. ruderatus and B. dahlbomii, or other hymenopterans were found. In farms stocked with commercial bumblebee colonies B. terrestris workers were also observed. The abundance of wild B. terrestris queens was positively associated with the area of natural forest and negatively associated with areas of high-food resources. These associations were significant at various spatial scales and peaked at 1 and 3.5 km radii for forest and high-food resources respectively . No correlation was found between natural forest area and high-food resources area at any spatial scale. However positive correlation of the proportion of natural forest and high-food resources areas between similar spatial scales was found. Among wild pollinators, B. terrestris queens were most abundant reflecting a successful spread of this exotic species in the study area. The only native bumblebee was almost absent which is consistent by previous reports on the decline of this species after the introduction of B. ruderatus and B. terrestrisin 1982-1983 and 1997-1998, respectively . The positive relationship between natural forest area and bumblebee abundance might reflect nesting suitability and/or hibernation habitat requirements of queen bumblebees. Undisturbed areas such as forests and forest edges can be suitable nesting habitat for bumblebees . These areas might be a source of continuous pollen and nectar resources throughout the foraging stage of bumblebees. While food supply in early spring can favor bumblebee colony establishment and initial growth , the reproductive success of the colony seems to be determined by late-season food availability provided by surrounding natural areas .. Late-season food supply is crucial for hibernating queens since Beekman et al. found that body weight at the start of diapause positively affects its success, while environmental temperature has no effect. Thus, forests can favor the number of queens produced per colony and/or favor hibernation success by providing food supply for queens before they start diapause. Further studies should focus on the relevance of food provision by natural forests and the type and quality of nesting/overwintering sites for bumblebees in south-central Chile. The results of this study also suggest that the presence of nearby flowering crops used by these pollinators can decrease the abundance of queen bumblebees in blueberry fields. This negative effect of high-food-resources areas could correspond to a ‘transient dilution effect’ in which the bumblebee density decreases as they are spread across larger areas of flowering crops and is consistent with previous studies that document such effect on oilseed rape . Simultaneously flowering areas probably compete for highly mobile pollinators such as bumblebees when they are available within the foraging range of the individuals.