We see that parcels receiving water deliveries are much more likely to plant strawberries or nursery crops at all levels of salinity, and are slightly less likely to plant vegetables. Anecdotally, when talking with growers in the Pajaro Valley, they stated that the recycled water is allowing them to grow strawberries in locations where the water quality was previously too poor. While this anecdotal evidence is clearly encouraging, I test this using the panel mixed logit structure by running a simulation that looks at how crop choices change for the parcels receiving delivered water if those deliveries no longer existed. I simulate and compute the annual utility-maximizing crop choices for each parcel under a scenario where there are no direct recycled water deliveries. To do this, I keep the estimated marginal utilities for the attributes of crops, parcels, and climate variables the same as in the panel mixed logit model in Table 3.3. For each parcel that currently receives recycled water, I estimate the probability of choosing each crop type and use these to predict each parcel’s baseline crop choice. Then, the simulation recalculates the probabilities of each crop being grown for these parcels after removing the water deliveries to reflect the no recycled water scenario. The differences in the estimated crop choice distribution is plotted in Figure 3.9, depicting which crops are chosen in the face of an elimination of recycled water. There is a dramatic increase in the amount of fallow ground,vertical gardening in greenhouse and significant declines in strawberry acreage and vegetables.
There is an increase in nursery crops, which are the most salt tolerant. Caneberries and orchards are not frequently planted in the delivered water zone, and experience virtually no change.The crop choice model above estimates the direct benefits to growers receiving recycled water deliveries, determining if growers are able to grow salt-sensitive, higher value crops. However, the model cannot capture the impacts of the recycled water on the underlying water quality. These impacts are paramount to understanding the effect recycled water has on mitigating seawater intrusion, and that benefits that may be gleaned by other users of the groundwater supply. I estimate these impacts in three ways. First, I conduct a simple, parcel-specific fixed effects regression, using the 2009-2020 data used in the same analysis as the crop choice model. Next, I use the extended water quality, recycled water, and water pumping data from 2003-2020 to estimate a staggered differences-in-differences model, where treated parcels are those that start receiving regular deliveries of recycled water. Finally, I implement an event study framework to evaluate how recycled water impacts water quality over time. To evaluate the indirect effects of the recycled water program, I see how recycled water deliveries impact the quality of the underlying groundwater aquifer. First, I look at an exploratory fixed effects model, using largely the same structure and data as the crop choice model. All estimates can be found in Table 3.5. Generally, the significance and coefficients on the control variables are in line with expectations. Of note, using more pumped groundwater does not have a significant impact on an individual parcel’s groundwater quality, which is important for the identifying assumptions of the crop choice model: growers do not have a significant individual impact on their own water quality.
The coefficients on the lagged delivered water are highly similar, but the estimation with the parcel fixed effects have higher standard errors. I weakly find that a one acre-ft increase in recycled water lowers groundwater salinity by 0.7 mg/L. While the signs on the delivered water coefficients are encouraging, this is unlikely to be enough to make any sort of sweeping claim on the effectiveness of recycled water, much less use these estimates in further analysis of the cost effectiveness of the recycled water program. Therefore, we turn towards staggered differences-in-differences and event study methods.The results from the difference-in-differences specification are presented in Table 3.6. Results suggest that a parcel that receives delivered water will see a 218.9 mg/L improvement in their water quality. This is a significant improvement, corresponding to approximately a 10% yield improvement in strawberries , or a 5% yield improvement in orchard crops . This kind of improvement in water quality is both impressive and realistic, as salinity levels do vary that much from the beginning to the end of the growing season for many parcels. While these results are encouraging, it is important to compare these results with the event study, to combat the concerns about the interpretation of staggered difference-in-differences methods. In addition, it allows us to see where the improvements are occurring in time, and whether or not they stay consistent. Results of the event study specification outlined in Equation 3.9 are shown in Figure 3.11. The overall treatment effect suggests that water quality improves by -239 mg/L after receiving recycled water deliveries. This is a similar magnitude as the DiD results, which had an overall treatment effect of -219 mg/L, and suggests that the staggered DiD results may not be overly biased.
It is also useful to note the pre-trends in the event study, which are insignificant from zero, and are also an important identification requirement for the difference-in-differences estimates. Perhaps the most striking results of the event study are the significantly negative spikes in years 7-9, as well as in year 15. On first glance, it is not immediately intuitive why treatment effects would be zero for the first six years, and then have a highly significant effect in improving water quality in year seven. However, when we look specifically at treatment timing and salinity levels in the region, the results fall into place. The median time in which an eventually treated parcel receives its first delivery is 2006. This means that on average, a parcel reaches treatment year seven in 2013, which corresponds to a massive increase in salinity in the delivered water zone, as shown in Figure 3.6. Scientifically, this makes sense: recycled water is going to have a bigger impact in improving water quality in years when the groundwater salinity is exceptionally high, especially when the quantity of recycled water is a much smaller fraction than the water in the groundwater basin. Moreover, the average TDS of the recycled water is frequently higher than the average TDS level in the delivered zone, as depicted by the dashed line in Figure 3.6. In fact, the only years in which recycled water quality is much higher than the average aquifer quality is in 2012-2013 and 2020. Therefore, we would expect that recycled water would significantly improve underlying water quality in those years, which is what is reflected in the event study figure. The results for the neighboring parcels are presented in Figure 3.12. This figure reports similar, but attenuated, findings to the water quality results directly underneath treated parcels. Across the board, confidence intervals are a bit noisier, but pre-trends are still at zero. The overall treatment effect is no longer significant, but water quality improvements up to 570 mg/L are reported in years 7-9 after the first recycled water deliveries. These effects are arguably the most interesting when considering the indirect benefits of the recycled water program. Although growers that receive recycled water deliveries still use some water from the underlying aquifer,greenhouse vertical farming growers without delivered water are the ones truly reaping the benefits from increased aquifer water quality. I use the event study results for the neighboring parcels to make a back of the envelope calculation of the benefits of recycled water on the aquifer. This allows for the most conservative estimate: there is no “double-counting” of the benefits for producers who receive recycled water, and who may not use as much groundwater as those without recycled water supplies. In years of high salinity, we see TDS levels substantially improve for neighboring parcels, up to 570 mg/L. According to the estimates, a 500 mg/L reduction in TDS would translate to a 16% increase in vegetable yields or a 39% increase in strawberry yields.
The WTP estimates calculated in the crop choice model say that this may have an impact of up to $123,000 per acre for vegetable producers and $86,850 per acre for strawberries. In crop revenue terms, a TDS improvement of 500 mg/L would have an impact of $14,345 on strawberries and $1,125 for vegetables per acre. For these neighboring parcels, benefits would add up to additional $10.78 million in benefits for strawberries and $1.52 million for vegetables. Of note, these are benefits that only accrue in years of especially high salinity. Building a recycled water facility and distribution system from scratch is not a cheap undertaking. In the Pajaro Valley, the recycled water facility cost $24.5 million USD to build, and the construction of the delivery system has brought the total up to $48 million. Additionally, PVWMA faces $12 million in annual operating costs, which largely goes to facility operation, improvements, and maintenance, although this also includes the costs of personnel and basin monitoring. These annual costs are offset by the $12 million in revenue generated by the groundwater pumping fees and recycled water delivery fees. In the direct benefits section, I find that without recycled water, producers would experience an average loss of $1,867 per acre. Largely, this seems like a reasonable estimate. On average, strawberry profits can reach up to around $12,000 in a “good” year. If strawberry yields were to be reduced by about 250 trays/acre , that would correspond to a profit loss of approximately $2000. In total, these direct benefits from the recycled water were estimated to average $16 million annually, which is substantially higher than the annual water management agency costs. I also find that in high salinity years, parcels neighboring those receiving recycled water could face benefits of $10.78 million for strawberries and $1.52 million for vegetables. These benefits are especially important because they reduce the negative impacts of extreme drought and salinity. These calculated benefits are short-run gains. However, there are also longer-term benefits that I cannot capture in this modelling framework: the reduction of seawater intrusion risk in areas that are not currently experiencing salinity issues. The added water to the basin and the reduction in pumping along the coast means that the eastern Pajaro Valley is less likely to experience seawater intrusion in the near future. While these benefits are difficult to estimate, this is likely to have impacts on long-term land values and crop yields. In sum, recycled water has positive benefits to agricultural production in this region for three primary reasons. First, the value of the agricultural produce is significant: profits can reach $12,000 an acre, and crop revenues up to $80,000. Second, the crops grown in the region are very salt sensitive: a 168 mg/L increase in TDS can decrease strawberry yields by 10%. Finally, the Pajaro Valley experiences very high salinity conditions, such that the relatively saline recycled water is of a higher quality than seawater intruded wells. Due to the combination of these factors, a small-scale, relatively expensive recycled water facility provides substantial benefits to the region. For recycled water to be beneficial in other locations, a combination of similar factors will likely need to apply. However, there are some additional attributes that could reduce the overall costs of recycled water. Regions with larger neighboring municipalities may be able to provide higher quantities of recycled water at a lower cost. Furthermore, areas with canals or pipelines already in place, such as regions using surface water for irrigation, may be able to implement a recycled water delivery system with reduced infrastructure costs.Coastal agricultural regions are facing several key issues under climate change: rising temperatures, increased precipitation variability, sea level rise, and a higher incidence of drought and extreme weather events. The interaction of these problems with limited groundwater supplies leads to the over-extraction of groundwater and seawater intrusion. Treated municipal wastewater is a promising strategy that can be used to prevent further groundwater overdraft and mitigate current damages from salinity. Through its use as an alternative water supply, growers can directly use this water on their fields, reducing their reliance on constrained groundwater resources and preventing further seawater intrusion.