While Northern California receives considerable seasonal rainfall, there is also significant spatial variability in rainfall totals and in corresponding summer flow persistence of small streams . Considering the ephemeral nature of surface water in many areas , the increasing frequency of drought due to climate change and cannabis cultivation’s consistent demand for irrigation water as crops near harvest , cultivators are strongly motivated to secure reliable water sources for the entirety of the growing season. Therefore, it is likely that water extraction from wells is a common practice for cultivators, beyond those seeking participation in the regulated industry . Although cannabis regulations place no explicit restrictions on where water is sourced, those currently within or seeking to join the regulated cannabis industry will be subject to a restriction on diversions of spring and surface water during the growing season . This requirement is already in place for permits issued by the California Department of Fish and Wildlife and will also be enforced by the State Water Resources Control Board beginning in 2019. The data provided in this study indicates that, in order to meet the forbearance period requirement, cultivators may be more inclined to drill a well to achieve compliance than to develop water storage for spring and surface water. Determining cultivators’ capability to store the water they need for the growing season may shed further light on the likelihood that growers will seek subsurface water. If compliance necessitates drilling a well,growers solution it will be important to account for the impacts of this potential shift in cultivation practices. Successful protection of freshwater resources in Northern California will require a more complete accounting of where cannabis cultivators source their water and the amount and timing of water extracted.
Study of cannabis as an agricultural crop has been notoriously inadequate, but data provided by the water quality control board’s cannabis program offers critical new insights into the water use practices of cultivators entering the regulated industry. In this initial analysis, we found that subsurface water may be much more commonly used in cannabis cultivation than previously supposed. Further analyses of cannabis cultivation’s water extraction demand, as well as of geospatial variation in water demand, may help elaborate the ramifications of this finding. Ultimately, a better understanding of cannabis cultivation’s water demand will be useful for placing the cannabis industry in the greater context of all water allocation needs in the North Coast and throughout California.Low-birth-weight infants are at high risk for morbidity and mortality in infancy and childhood as well as chronic diseases in adulthood . More than 95% of LBW infants are born in low-income countries . In 2002, the World Health Organization estimated that 12% of infants born in Guatemala were LBW [Pan American Health Organization 2007], although this is likely to be an underestimation of national rates, because those most likely to be LBW and least likely to be weighed at birth are rural, poor, indigenous children born at home. Solid fuels, including coal, wood, crop residues, and animal dung, are used by about half the world’s population and by 90% of people residing in rural areas of low-income countries . Traditional cooking stoves are typically inefficient at combusting solid fuel and often lead to high exposures to particulate matter and carbon monoxide . In Guatemala, where 83% of the rural households use wood for cooking fuel , 24-hr kitchen concentrations of PM2.5 from wood fires can range from 100 μg/m3 in homes using chimney stoves in good condition to 1,000 μg/m3 in homes using open fires. These concentrations are orders of magnitude higher than those in homes where tobacco cigarettes are smoked, where daily mean concentrations of PM2.5 vary between 25 and 40 μg/m3 . Mean concentrations of CO over a 24-hr period are also high and often exceed the health-based WHO guidelines of 9 ppm .
Depending on source strength, ventilation, and other factors that mediate or magnify exposure, wood fires typically yield levels of exposure to PM and CO that are much higher than environmental tobacco smoke . ETS has been linked to mean birth weight reductions of 33 g [95% confidence interval , 16–51 g] and to a 22% increase in the prevalence of LBW infants . Associations between maternal exposures to household air pollution and LBW have been described in children in Guatemala , Zimbabwe , Pakistan , and India . In previous studies, all exposures to solid fuel smoke were estimated from self-report of fuel and stove type used. Thus, the level of exposures to specific household air pollutants is not known. The purpose of this study was to determine the impact of reduced HAP from an introduced chimney stove on the prevalence of LBW infants in rural Guatemala. We assessed exposure with personal measures of CO during pregnancy. For this study, we conducted a post hoc analysis to examine the birth weight of Guatemalan infants whose mothers participated in the RESPIRE trial. The primary aim of RESPIRE was to examine the incidence of acute lower respiratory infection among children < 18 months of age in households randomized to receive a wood-fueled chimney-stove intervention compared with control families who continued to use traditional open fires .Between October 2002 and December 2004, the RESPIRE trial was conducted among 534 households in 23 rural Guatemalan communities. Most participants were members of the Mam-Mayan linguistic group. Households were recruited if they used open wood fires for cooking and had either a pregnant women or a child < 4 months of age.The other half of the homes continued to use an open fire until their child reached 18 months, or until the study ended, at which time the family was offered a chimney stove.
A total of 266 of the 534 women were enrolled during pregnancy, with the remaining children recruited at < 4 months of age. After enrollment, pregnant women received a baseline pregnancy examination by a physician employed by the study, which included a medical history, urine analysis, blood pressure measurement, and estimate of fetal growth . Fieldworkers visited women weekly in their homes, and women were referred to physicians if the women reported pregnancy complications. Among the enrolled pregnant women, 254 women delivered singleton, live births . Twelve additional births occurred among these women during the study period. Figure 1 summarizes characteristics of pregnant women and births. Among the 254 women, 240 participated in an ancillary study of adult women’s lung function. In this study, height and non-pregnant weight were measured . Height was measured using a seca portable stadiometer . Weight was measured using a Taylor Precision Tech Lithium electronic scale . Once RESPIRE staff were notified of a birth, either by a fieldworker at the weekly home visit or by a family member who contacted project staff, the project manager attempted to visit the home to weigh the newborn within 48 hr. Sometimes practical issues made it difficult to reach the home to carry out measurement, such as delayed notification by family members,growers equipment minimally skilled birth attendants who did not routinely weigh newborns, and communities located several hours away from the project office. Thus, of the 254 newborns, 208 birth weights were measured within 1 week, including 174 measured within the first 48 hr. Because newborns can lose up to 5–10% of their birth weight in the first week, we restricted our birth weight analysis to the 174 newborns measured within 48 hr of birth. Birth weight was measured in grams using a calibrated Siltec BS1 baby scale with 10-g readability . Infants wore a light shirt provided by project staff to standardize clothing weight. Gestational age was not assessed. Intention-to-treat analysis assumes that those who were randomized to receive the intervention stove had lower exposure to HAP than did those in the control group who continued to use open fires. However, because the RESPIRE trial was designed to monitor infant respiratory illness, many of the pregnant women were recruited in the later stages of pregnancy. They received the stove, which required a 5-week drying period after construction, shortly before, or even a few days after, delivery. Therefore, birth weight outcome analyses were based on actual stove type in use during the pregnancy during the observation period and not based on randomization, resulting in 69 births to women who used a chimney stove during pregnancy and 105 births to women who used an open fire during pregnancy .
Personal exposure to CO was used as an indicator of exposure to HAP and as a proxy for PM exposures . Passive-diffusion colorimetric CO tubes were used, with high-range tubes for baseline period measurements before stove construction, and low-range tubes for observation period measurements. Methods for deploying, validating, and analyzing the CO tubes have been described previously . Pregnant women wore CO tubes for a 48-hr period, with a total of 378 CO tubes worn during 247 pregnancies. Baseline measurements were made before stove construction when all houses were using open fires. During this period, women wore 278 CO tubes during their pregnancy, with 4% of women in the first, 33% in the second, and 63% in the third trimester. During the observation period, CO measurements were made at 3- to 6-month intervals. Pregnant women wore 100 CO tubes during this period, with 6% of women in the first, 14% in the second, and 80% in the third trimesters. If a woman had more than one measurement taken during either the baseline or the observation period, an average of the measures was provided for that period. However, with so few measures available for pregnant women in the first two trimesters with the stove in use, we did not have enough information to incorporate these data into the multivariate birth weight models. Instead, the independent variable is actual stove type used during pregnancy. Statistical methods. We used Stata version 11 for data analyses. Because CO tube data had a right-skewed distribution, both arithmetic and geometric means and respective standard deviations are reported and log-transformed data were used to test for differences between groups. We compared stove groups and birth weight outcomes with Student’s t-test for continuous variables, and chi-square tests of significance for categorical variables. We constructed multivariate linear models of birth weight as a continuous outcome among exposed and unexposed women, based on the stove type that was used during the pregnancy. Covariates were included in the multivariate linear regression if a) there was evidence of a relationship with either outcome or HAP exposure in the literature, or b) there was an observed statistical association with either outcome or exposure using appropriate parametric or non-parametric hypothesis tests. To investigate potential seasonal effects of HAP on birth weight , we examined season of birth in the multivariate model. The three seasons in this region of Guatemala are cold season , warm, dry season , and warm, rainy season . We also examined the effect of season during the three trimesters of fetal growth; interaction terms for season × trimester of stove introduction were included in the regression models. Variables in the model were assessed for multi-collinearity using a variance inflation factor and were removed if VIF > 10. Outliers were classified using studentized residuals ± 3.0, and a model without outliers was compared with a model that included the “outlier” observations. Leverage plots and partial regression plots were examined to assess any influence individual observations might have on model coefficients. Predicted residuals were plotted against a normal distribution and against the fitted values, to assess for functional form and constant variance. We used the likelihood-ratio test statistic to compare restricted versus full models to determine which model best described the data. We clustered multiple observations on the same woman to derive standard error estimates that would be robust to nonindependent observations. In the final multivariate model, two outlier observations were excluded because they were found to sufficiently influence the homogeneity of the residuals of the data. An adjusted odds ratio for presence of LBW among open-fire users, compared with chimney-stove users , was calculated using logistic regression, and CIs were estimated using robust error estimates. p-Values < 0.05 were considered significant.