Long-term predictions of the water cycle in a watershed must account for the human factor. Agriculture is by far the largest water user,making up 70% of water use worldwideand over 95% of water use in Montana. Predicting agricultural water use is complicated because farmers make decisions based on climatic conditions, economic conditions, and personal preferences. Coupled models that account for both climate and agricultural economics are necessary to understand and predict the whole water cycle.

The impact of agriculture at the watershed scale is effectively determined by total agricultural area and crop water requirements, and ultimately comes to bear on the flows at the watershed outlet. In many watersheds total agricultural area is declining. As urban areas grow, the market value of agricultural land increases. If the value of a Farmer's land is higher than the expected return from the sale of crops, the land will likely be sold and divided into smaller parcels. This sell-off is influenced by climate conditions, which affect crop yields, and by economic conditions, such as crop prices or input costs. Accounting for these factors at the scale of individual farms is impractical, but at the scale of a watershed the sell-off of parcels could be statistically modeled to capture the overall behavior of the agricultural system. However, the agricultural parcel sizes of a watershed have a complicated distribution that is not easily represented statistically.

We have developed a method for representing agricultural land as a mixture of Gaussiandistributions. Because the individual Gaussian distributions are symmetric, they can be propagated through linear models of agricultural economics that calculate the fraction of agricultural land for which market value exceeds expected returns. By assuming that this land is sold and ultimately converted to urban use, we can calibrate this model to reproduce an observed loss of agricultural land area over time. We have applied our model to the Bitterroot River Watershed, which is an ideal case study due to its strong agricultural economy and steady urban growth. We have been able to reproduce observed trends in agricultural land area, including accelerated sell-off during periods of prolonged dry weather. We use calibrated model to estimate the sensitivity of the system to climate scenarios, such as drought, and economic scenarios, such as reduced crop prices. The model is also used for stability analysis – under steady conditions the coupled models converge to an equilibrium point – highlighting potentially unsustainable combinations of climatic and economic conditions. These results further our ability to manage land and water resources in a way that can maintain both healthy agricultural economies and healthy flows in our rivers.

The long-term effects of pre-commercial thinning on carbon storage and distribution in western larch (Larix occidentalis) stands of the Northern Rockies: Insights from a long-term silvicultural experiment.

Principle investigator: Michael Schaedel

Using forests to sequester and store carbon in order to mitigate climate change has gained interest in public policy discussions and is increasingly becoming a land management objective. Research indicates that forests currently sequester approximately 10% of annual US fossil fuel emissions and it is suggested that through intentional carbon management that this rate could be increased. Forests store atmospheric carbon in biomass through photosynthetic uptake. Many studies show that high tree densities may store the greatest amount of on-sight carbon. However, there are trade-offs to managing forests solely for carbon sequestration, namely a decrease in forest resilience that may lead to a higher frequency and severity of fires, insect outbreaks, and disease. These large scale disturbances cause the release of much of the stored carbon in the forest ecosystem. Tree density management, often accomplished through thinning, is one of the major tools foresters employ in order to improve forest health and resilience. While thinning is a common practice in restoring resilience to forest ecosystems and managing forests for the production of wood products little is known about the long-term effects of thinning on the ability of forest to act as carbon sinks. This study examines how the common practice of forest thinning affects carbon storage in western larch dominated mixed-conifer forests in the Northern Rockies.

Past studies of the effects of thinning and partial harvest on carbon storage have contradictory findings; some studies show that thinning causes increased carbon storage, attributable to increased growth rates of residual trees, while other studies find a decrease in carbon storage due to the removal of tree biomass from the forest stand. These conflicting results are potentially caused by several differences in study designs: 1) thinning objectives, 2) carbon storage pools measured, 3) length of the study and, 4) age of the stand when thinned. The five major pools of carbon storage are live-tree biomass, live-understory biomass, dead woody detritus, forest floor litter and duff, and mineral soil.

This study is superimposed on a long-term western larch spacing study established in northern Montana in 1961. The larch spacing study used a 3x3 factorial design made up of three thinning intensities (residual tree densities of 494 TPH, 890 TPH, and 1640 TPH) and three thinning intervals ( entries every 10 years (4 entries), every 20 years (2 entries), and every 40 years (1 entry)). Data was collected at a five year intervals from 1961 to 2001 on a suite of attributes of tree growth and vigor. For this study, carbon will be measure in five major pools: live tree, understory vegetation, woody detritus and snags, forest floor, and mineral soil. To capture live-tree biomass a re-measurement of the larch spacing study will be completed to capture tree attributes on all of the 0.04 ha treatment plots. Allometric equations will be used to convert tree volume to carbon biomass. Understory vegetation will be sampled in 0.25 subplots by clipping the vegetation and taking it back to a lab to analyze carbon content. Dead woody detritus will be measured by doing a census of woody debris in the 0.04 ha plot and converting the volume to carbon using established equations that take into account changes in wood density due to the stages of woody decay. The forest floor will be sampled in 0.05 circular plots where all organic material is taken to a lab and burned to analyze carbon content. Mineral soil will be sampled using soil cores taken to a depth of 30cm and analyzed by burning the soil to measure the carbon content. Measurements will be taken in the summer of 2015. Live tree and understory carbon will be tracked through time using the new data in conjunction with the data from the long-term western larch spacing study.

An oral presentation of this study will include an analysis of live tree and understory carbon storage trends through time utilizing existing data from the larch spacing study. By measuring 5 major carbon pools and using a long- term data set this study will be able to provide more conclusive information on the effects of thinning on carbon storage. This data will help land managers identify effective ways of managing mixed-conifer forests as carbon sinks and asses any carbon storage trade-offs made while managing for other objectives, such as forest resilience and wood production.

Rationale: Particulate matter (PM) is a measurable component of air pollution that has been associated with adverse cardiovascular and respiratory outcomes. Research indicates environmental factors such as air pollution are involved in changes through epigenetic mechanisms during development that may persist into adulthood and even span multiple generations of inheritance. Epigenetics is the study of heritable changes of gene expression that do not alter the actual DNA sequence. One epigenetic mechanism is DNA methylation. Long Interspersed Nuclear Element (LINE-1) is a DNA repetitive element that can be used as a proxy measurement of DNA global methylation.

The purpose of this pilot study was to compare epigenetic biomarkers across different sample matrices (i.e. blood and buccal) and across related subjects (i.e. maternal and infant).

Methods: Informed consent was provided by pregnant women (n=23) who were recruited through Women, Infants, and Children (WIC), hospital birthing classes, or flyers in obstetrician’s offices. Demographic and medical data was collected from hospital records for both mothers and newborns after birth. Follow-up health surveys were administered by telephone that were designed to collect indicators of pre-asthmatic respiratory symptoms or conditions.

Biological samples were collected before or shortly after time of birth at Community Medical Center of Missoula, MT. The samples collected were maternal blood (n=15), umbilical cord blood (n=15), and maternal (n=23) and newborn (n=23) buccal (cheek) cells. Buccal cells were collected and processed according to the Gentra Puregene Kit (Qiagen, Germantown, MD). These biologically accessible tissues serve as surrogates to study gene methylation associated with respiratory health.

Samples were stored at -80°C until DNA extraction and subsequent bisulfite treatment. The samples were amplified in duplicates with polymerase chain reaction (PCR). LINE-1 methylation was analyzed with pyrosequencing on a Pyromark Q96 MD (Qiagen, Germantown, MD). All statistical analysis was performed in Statistical Analysis Software (SAS, version 9.3).

Results: The mean (standard deviation (sd)) of LINE-1 methylation percentage for mother and infant buccal cell derived DNA were 58.75 (3.89) and 57.16 (2.54), respectively. Percent methylation maximum for mother and infant buccal samples were 70.22 and 64.25, respectively, and minimum were 54.86 and 52.94, respectively. Paired t-test indicated that LINE-1 methylation percentages in maternal buccal samples were higher than methylation percentages in the paired infant samples (mean difference (95%CL) = 4.4 (2.3, 6.6)).

The mean (sd) of LINE-1 methylation percentage for mother and infant/cord blood derived DNA were 75.19 (3.17) and 75.86 (3.05), respectively. Percent methylation maximum for mother and infant blood samples were 79.42 and 79.50, respectively, and minimum were 70.39 and 69.31, respectively. Paired t-test indicated that LINE-1 methylation percentages in maternal blood samples were similar to methylation percentages in infant blood samples (mean difference (95% CL) = 0.66 (-2.0,3.3)).

Conclusions: LINE-1 methylation percentages between sample matrices (i.e. blood and buccal) and subjects (i.e. maternal and infant) were not correlated. The percent methylation of LINE-1 in DNA from blood was consistently greater than for DNA from buccal tissue for both mother and newborn samples. It was expected that LINE-1 measurements for blood DNA would differ from buccal DNA because circulating blood represents a more diverse cell population. Gene-specific methylation of the promoter region for interferon-γ, a cytokine associated with asthma, will be studied with the remaining samples of bisulfite-treated DNA from this study. Epigenetic changes may serve as useful biomarkers for predicting asthma risk in children exposed to biomass smoke. These methods can be applied to future studies to investigate the epigenetic relationship of prenatal asthma risk and PM wood smoke exposure.