We developed a new experimental system with which anelasticity of a rock analogue sample (polycrystalline borneol) can be measured in situ during dislocation creep. We attached a piezo-electric actuator to a triaxial deformation apparatus, to add a small cyclic load to a large constant load. We also attached a load cell and two laser displacement meters to measure the small cyclic load and displacement accurately. Using this new system, we deformed a polycrystalline borneol sample under diffusion creep (σ = 0.27 MPa) for about 1 day, intermediate creep (σ ≈ 1.1 MPa) for about 0.5 day and dislocation creep (σ ≈ 2.2 MPa) for about 1 day continuously, and performed an in-situ measurement of Young’s modulus and attenuation at frequencies of 2.5 and 1.0 Hz. During the first diffusion creep, Young’s modulus increased, probably due to the improved contact between the sample and the piston, and reached a constant value. Although the modulus did not change during the second intermediate creep, it gradually decreased during the third dislocation creep. The final modulus reduction was about 20%. The present result supports our previous result obtained from the anelasticity measurement of a pre-deformed sample (Sasaki et al., 2017, AGU fall meeting; Sasaki et al., 2019, submitted to JGR). In the previous study, we showed that detailed form of the dislocation creep curves can be explained well by considering a gradual increase in dislocation density during the dislocation creep. Also, reduction of Young’s modulus by about 10% was observed after the dislocation creep under the similar condition to this study. The gradual decrease in Young’s modulus observed in this present study provides a supporting evidence for these previous results. Frequency dependence of the dislocation-induced anelasticity obtained from our previous study is a peak at much higher frequency than the grain boundary-induced anelasticity (Sasaki et al., 2017, AGU fall meeting; Sasaki et al., 2019, submitted to JGR). In contrast, dislocation-induced anelasticity obtained from single crystal forsterite or polycrystalline olivine is a broad absorption band (Guéguen et al., 1989; Farla et al., 2012). It is important to clarify the reason for this discrepancy.

Wetlands are the most important natural source of methane (CH4) to the atmosphere. When present, plant efflux is often the dominant pathway for methane emissions from the anoxic wetland sediments, however, they are rarely considered in chamber measurement campaigns of CH4 flux in wetlands. While studies have focused on the mechanisms that govern transport through plant structures, none have addressed how CH4 emissions relate with carbon uptake or how this relationship may vary among plants. Understanding this relationship has important implications for linking carbon flux at the ecosystem level to wetland CH4 emission predictions, and ultimately, to help improve current carbon budget and greenhouse gas models of wetlands. This study presents simultaneous measurements of CH4 and CO2 fluxes in three macrophyte species of different cosmopolitan genera: Typha (emergent), Nelumbo, and Nymphaea (floating attached). These measurements were conducted using the Picarro GasScouter G4301 fitted with custom-designed leaf chambers. Chamber measurements were performed three times during a day (morning, noon and afternoon), in three separate months. Results indicate significant negative correlations between CH4 flux and carbon uptake in Nelumbo and Nymphaea species and a non-significant positive correlation in Typha sp. We term the rate of CH4 flux per unit CO2 uptake – CH4 flux efficiency. We found that in Nymphaea CH4 flux efficiency is almost double that of Nelumbo. These results suggest the dependence of CH4 efflux in stomatal conductance in the floating attached macrophytes, while in the Typha CH4 flux is primarily through other pores, probably aerenchyma, not regulated by stomata. Our results emphasize the importance of species-specific knowledge and considering the growth type of different wetland macrophytes (e.g. emergent vs floating attached) in scaling of bottom-up observations and ecosystem models.

One third of all coastlines worldwide consist of permafrost. Many of these permafrost coasts are presently exposed to greater environmental forcing as a consequence of climate change, such as a lengthening of the open water season, intensified storms, and higher water and air temperatures. As a result, increasing erosion rates are currently reported from various sites across the Arctic. It is crucial to synthetize these data on Arctic shoreline dynamics in order to improve our understanding on present coastal dynamics on the pan-Arctic scale. A first synthesis product was released in form of the Arctic Coastal Dynamics databse in 2012, which included data published until 2009 (Lantuit et al., 2012). Since then, numerous publications and data products were published on short and long term changes of Arctic coasts across a wide range of study sites. We made an extensive literature review of publications released within the last 10 years and updated the shoreline change data section in the Arctic Coastal Dynamics database. While in 2009 for one percent of the Arctic shoreline data on coastal dynamics was available, the addition of new data leads to a broader data coverage, which is mainly the effect of the greater availability of remotely sensed products for analyses conducted in these remote regions. Further, the additional data allow us to update the current mean rate of Arctic shoreline change.

The intraseasonal variation of the Tibetan Plateau summer monsoon (TPSM) during 1979–2011 is investigated. The TPSM shows a dominant quasi-biweekly oscillation (QBWO) in most summer seasons, and its active/break phases are closely related to more/less precipitation over the Tibetan Plateau. We suggest that the TPSM QBWO is associated with a southeastward propagating nonstationary wave train in the middle and upper troposphere. It shows equivalent barotropic vertical structures over the midlatitudes and a baroclinic structure over the eastern Tibetan Plateau. Wave activity flux analysis indicates that it originates from northern Europe, which is an active center of the summertime Arctic Oscillation (AO). The AO also shows significant QBWO signals and leads TPSM QBWO by about 13 days. Phase composite and wave activity flux analyses of AO QBWO confirmed that the wave train influences TPSM QBWO, suggesting that AO plays an important role in the TPSM on a 10- to 20-day timescale.

Extreme floods which overwhelm the capacity of a system of flood control dams may result in overtopping one or more of those structures. Traditional US Army Corps of Engineers analysis of hydrologic hazards isolates the study area to a single dam. However, in watersheds where flood hazard is managed by several dams, the estimate for the annual probability of overtopping a dam may be influenced by the operation of one or more other dams in that system. Evaluation and prioritization of modifications for dam safety in a portfolio of structures requires a sound estimate of overtopping probability for every structure. In an effort to properly characterize the hydrologic hazard for five dams in the Trinity River Basin above Dallas, Texas, synthetic weather generation coupled with hydrologic and reservoir models is applied to extend the stage-frequency curve for each dam beyond the observed record. The synthetic weather model is comprised of processes which typify floods most likely to result in overtopping the study dams: 1) continuous, local-scale precipitation and temperature sampling to characterize antecedent hydrologic conditions, 2) intermittent (inhomogenous Poisson), synoptic-scale precipitation sampling based on regional precipitation-frequency analysis to generate hazardous floods, 3) k-nearest-neighbor resampling of precipitation and temperature spatiotemporal patterns and 4) temporal disaggregation of daily precipitation to hourly using correlated Brownian processes. Interrelations between local-scale precipitation, synoptic-scale precipitation and temperature are preserved using a Gaussian copula. Natural variability in annual maximum reservoir stage is described using a stratified sampling scheme used to disproportionately represent extreme floods in a fixed sample of 1,000 events, resulting in fewer model events required to span the probability space from 0.5 to 10-8 annual exceedance probability. Knowledge uncertainty in model components is estimated using a parametric bootstrap, resulting in multiple realizations of synthetic weather. Each weather realization of 1,000 events generated using varying parameters is routed using hydrologic and reservoir models for the system which produce a posterior distribution of annual overtopping probability for each structure.

Mongolia has a complex tectonic history. The lithosphere was formed from multiple plate collisions in the Neoproterozoic - Early Paleozoic associated with the Central Asian Orogenic Belt. The region has since been modified by Mesozoic rifting, Cenozoic magmatism, and major strike-slip faulting along terrane boundaries and sutures. Central and Western Mongolia are part of the larger high elevation, low-relief Mongolian Plateau. To gain deeper understanding of modern deformation within the Hangay Dome in Central Mongolia, two years of teleseismic, regional, and local seismicity, recorded by a dense array of 72 temporary broadband seismic stations, was used to determine the distribution of seismicity and crustal structure. Results from receiver function analysis indicate the Hangay Dome has a crustal thickness ranging from 41-59 km. The thickest crust resides under areas of high topography and generally thins to the east. Average Vp/Vs ratios range from 1.77-1.8. We located the 7680 events detected by the array using a local 1D velocity model. Many events outline the Bulnay and Bogd faults, where historic Mw 8 earthquakes have occurred. Considerable seismicity is observed on the South Hangay – Bayan Hongor Fault System, including a Mw 4.6 earthquake. Seismicity is also observed along the Egiin Davaa and Mogod Faults. Preliminary results from a joint tomographic inversion for earthquake location and 3D velocity structure show a relatively uniform crust, where P-wave velocities in the uppermost crust range from 5.8-6 km/s. In these preliminary inversions, large portions of the region show Vp exceeds 7.0 km/s in the lower 10-15 km of the crust. The depth to the Moho is consistent with results from the receiver function analysis. Lateral velocity variations generally align with terrane boundaries and faults, such as the South Hangay - Bayan Hongor Fault System. Seismicity relocated in the inversion outline the South Hangay, Egiin Davaa, and Bulnay Faults. In addition, a cluster of seismicity locates between the Egiin Davaa and Hag Nuur faults, where no fault has previously been mapped. Seismicity in the Hangay Dome is generally confined to the upper 20 km, suggesting a rheological transition from brittle to ductile at this depth.

The vertical distribution of plant roots in the soil profile is a key trait modulating plant contributions to soil carbon storage, drought and nutrient stress resistance, yield, and fitness. However, direct sampling of deep roots requires massive effort, so existing data are sparse and many researchers have adopted modeling approaches to fill data gaps and generate hypotheses about how soil properties change the biogeochemical, agricultural, ecological, and hydrologic consequences of root depth. Such models are useful only if they correctly represent the processes of interest and give accurate predictions of the root systems they simulate. Most current root growth models represent soil as a uniform and unrestrictive medium. This is often a reasonable simplification when modeling roots grown in pots or artificial media, but is less so for field soils which often increase in density, hardness, and heterogeneity with depth. To better predict the effect of soil hardness on root distribution, we updated the structural-functional root growth model OpenSimRoot to explicitly predict soil hardness from soil bulk density, water content, porosity, and depth. Root growth impedance is curently represented by linear scaling of the root elongation rate according to soil hardness. Future work will incorporate configurable growth responses and allow hardness to control changes in root diameter and growth direction, thus allowing the model to examine the fitness implications of carbon reallocation in complex structured soils. Our updated OpenSimRoot captured >50% of observed variation in penetrometer resistance from field soils. When we incorporated soil hardness into simulations of maize growth, we observed a substantial reduction in the predicted root:shoot ratio that overwhelmed previous model predictions of increased water uptake from steeper root angles. These findings reinforce that models considering costs and benefits of deep rooting should routinely consider soil hardess.

Rangelands cover over 50% of the land surface area in the western US, providing important economic, social and environmental benefits. The resilience of western rangelands, however, is threatened by climate change, including altered phenology and precipitation patterns, increased frequency and intensity of drought and forest fires, heightened pressure from invasive plants, and reduced water storage in winter snowpack. Climate adaptation strategies are available to ranchers, yet uptake varies substantially. Rancher decision-making is a complex function of their beliefs, knowledge, skill level, risk perceptions, and the institutions supporting them. Semi-structured interviews, focus groups and workshops will be utilized to examine how ranchers in Idaho, Montana, Wyoming perceive and respond to climate change, and the opportunities and barriers these social processes create for climate change adaptation.

An unanswered question in planetary science is how could the early Martian atmosphere have maintained a greenhouse effect sufficient to allow for liquid water on the surface? A recent study by Wordsworth et al. (DOI:10.1002/2016GL071766) suggested that previously unaccounted-for collision-induced absorption (CIA) by carbon dioxide (CO2) and hydrogen gas (H2), and by CO2 and methane (CH4) could provide the additional atmospheric absorption needed to trap enough radiation to raise the Martian surface temperature above freezing. However, as CIA cross-sections for CO2-H2 and CO2-CH4 complexes do not exist in the literature, the authors could only use computational methods to simulate the CIA absorption cross-sections that they themselves identify in the study as needing experimental validation. Preliminary results will be presented from experimental measurements of the CIA cross-sections for CO2-H2 and CO2-CH4 complexes performed using Fourier Transform Spectroscopy. We have obtained Beam-time at the Canadian Light Source Far-IR beamline in late October and early November which will allow us to derive Cross-sections over a spectral range of 0-3000 cm-1 and a temperature range of 200-350 K. In addition to allowing us to experimentally validate the hypothesis of Wordsworth, the cross-sections so obtained can also be applied to other planetary systems with CO2-rich atmospheres, such as Venus, and will be useful to terrestrial spectroscopists.

Forests are becoming drier due to a warming climate thus increasing the risk of wildfires. In recent years, wildfires have grown larger and more severe. In the U.S., over 80% of wildfires are human-caused and such events can substantially extend the fire season. At the same time, more and more people are living in areas where wildfires can burn. Recent fires have illustrated the devastating consequences of fires in the Wildland Urban Interface (WUI). In this context, the ability to rapidly assess fuel load is crucial in assessing and managing the risk of wildfires. Current methods for monitoring fuel loads (e.g., FIREMON, Brown’s Transect) are accurate but time- and personnel-intensive. The Global Learning and Observations to Benefit the Environment (GLOBE) Observer Fire Fuel app is being developed using the Photoload Sampling Technique to offer options in rapidly assessing fuel loads by 1) providing citizen scientists with a fast and easy method to monitor WUI fuel loads, enabling them to contribute to the knowledge of fuels in their communities, and empowering them to think more about how fuels might be managed in their area; and 2) offering natural resource managers and fire science researchers a detailed, scientific application that primarily aids experts already studying fuels to better collect the fuels data they need. This poster will provide an overview of the GLOBE Observer Fire Fuel app and the current app development status. We will highlight the value and opportunity the power of smartphones and tablets offer to rapidly assess fuel loads via an app-based method compared to collecting the data on paper. We greatly welcome input from the fire science community at this point of the fire fuel app development.

Human activities have been changing deposition rate of the atmospheric reactive N over the last decades. The understanding of the process that rules the accumulation and deposition of reactive nitrogen in the environment still faces major gaps mainly in regions with lack of data as South America. In this work, we evaluated the atmospheric dry deposition of reactive nitrogen (Nr) using concentration measurements of different chemical species, in six sampling sites with different environmental characteristics in the state of São Paulo, Brazil, for a sampling period of approximately 24 months (2015-2017). We used the sampling system DELTA (Sampler for Atmospheric Sampling of Long Term), and ion chromatography method to determine the concentration of gaseous (NH3 and HNO3), and particulate (NH4+ and NO3-) chemical species in the atmosphere. The N-NH3 (gaseous) was the dominant form of Nr in the atmosphere at all sampling sites. The highest N-NH3 mean value was found near the urban area of the municipality of São Paulo (SP, 1.58 ± 0.73 μgN m-3) and the lowest mean value in the most eastern sampling site (0.26 ± 0.26 μgN m-3). The high values in the SP sampling site is related to the intense traffic in the metropolitan area and, also to waste management, industrial NH3, and human emissions as observed in other urban areas around the world. The particulate N-NH4+ is the second most common form of Nr in the São Paulo state atmosphere. The other forms of Nr in the atmosphere, the gaseous HNO3 and particulate NO3, represent only about 10% of the total Nr in the atmosphere each. The total gaseous Nr concentration was a factor of 1.7 larger than particulate Nr. The Nr deposition varied throughout the years. We did not observe a pattern of variation linked to meteorological characteristics of dry and wet season, as observed in other regions of the globe. However, we found good correlation with wind speed higher than 3.5 m s-1 and humidity. Our results provide the first spatial analysis of Nr deposition using in situ data in a Latin American region and will contribute to the understanding of nitrogen balance and to improve Nr deposition modelling approaches. This study was supported by the project Nitrogen cycling in Latin America: drivers, impacts and vulnerabilities (Nnet, IAI/CRN3005 and FAPESP 2012/06416-1), PCI Program of the MCTIC, and collaborators.

When NASA established the Planetary Data System (PDS) in the late 1980s, its mandate to the PDS was not merely to preserve the bytes from NASA’s planetary science missions, but to maintain the usability of the data for present and future generations. Two fundamental pillars support this ambitious goal: The external peer review required for acceptance of all archived data submissions; and the PDS Standards for data and metadata formatting and completeness. The PDS external peer review process is at least equivalent to, if not more rigorous than, the journal refereeing process(1). Data reviewers who are field experts but not affiliated with the data preparer, nor involved in the PDS consulting process, are brought in to review documentation and completeness. They are specifically charged to attempt to use the data to perform some scientific investigation (reproducing published results, comparison to correlated observations for consistency, etc.). If the reviewers are not successful, the impediments are documented and the data submission is amended by the preparer until the reviewers are satisfied. This process demonstrates immediate usability of the data. The PDS Standards, and in particular the recently-implemented version based on the PDS4 Information Model, require exhaustive metadata documenting data structure, observing circumstances, provenance, analytical metadata, and so on using the same templates across the entire archive. The associated schematic enforcement of at least minimal requirements for metadata completeness and quality provides a foundation for discoverability, interoperability, and usability of data from disparate sources throughout the archive. Together, the PDS external peer review and the Information Model-based PDS4 standards ensure both quality and usability for data accepted into the PDS archive, for this and future generations of planetary scientists. Reference: (1) Raugh, A. and Bauer, J., PDS Data Sets as Peer-Reviewed References, Poster presented at the 15th Annual Meeting of the Asia Oceania Geosciences Society Meeting, 03-08 June 2018, Honolulu, Hawai’i.

Several studies suggest that magnetic reconnection plays an essential role to generate and accelerate most of the erupting coronal magnetic flux ropes such as coronal mass ejections (CMEs). We explore the connection between magnetic properties (magnetic flux and helicity) of interplanetary coronal mass ejection (ICME) flux-ropes (magnetic clouds [MCs]) and those of associated near-sun CME flux-ropes formed in situ by low corona magnetic reconnection. We identify the progenitor CMEs and their solar sources and derive the source region reconnection flux using the post-eruption arcade (PEA) method. Combining the reconnection flux and the geometrical properties of associated CMEs obtained by forward-modeling, we extract the magnetic properties of CME flux ropes at their source. To measure the magnetic properties of 1 AU ICME we use constant-α force-free cylindrical flux rope model fit to in situ observations and directly from the observed magnetic time series rotated to the cloud frame. We investigate whether a significant difference exists in magnetic properties of ICMEs if their solar source is composed of pre-existing flux-ropes (filaments). This study has significant implications in finding the role of reconnection in the formation of twisted flux ropes during a solar eruptive process that transport solar magnetic flux and helicity into interplanetary space.

We aim to better understand the spatial distribution of interseismic overriding plate deformation at and near subduction zones. To this end, we analyze horizontal GNSS velocities in South America, southeast Asia, and northern Japan, computing and interpolating local trench-normal and -parallel velocity components. Velocities generally decrease with distance from the trench with a steep gradient up to a “hurdle”, beyond which the gradient is distinctly lower and velocities are near-zero. The hurdle is located 500–1000 km away from the trench for the trench-perpendicular component and either at the same distance or closer for the trench-parallel. In contrast, significant displacements during large megathrust earthquakes are generally observed beyond the hurdle. To test our hypothesis that the hurdle results from a lateral contrast in overriding plate compliance, we use cyclic three-dimensional finite element models . Our results are consistent with the observed interseismic velocity gradients and far-field coseismic displacement. The gradient in modeled trench-perpendicular velocities depends on the location of the contrast and on the plate compliance on both sides. Trench-parallel velocities have a progressively shallower gradient with distance from the trench and only depend on the near-trench modulus. The inferred contrast probably results from thermal, compositional and thickness contrasts. This interpretation is consistent with the presence, close to the observed hurdle, of major tectonic or geological boundaries separating the plate margin from a distinct, and likely less compliant, plate interior. Stress accumulation on the model’s locked megathrust patches is hardly affected by the distance to the contrast.

Hillslopes are responsible for the production and transport of sediments within a landscape (Gilbert 1877). Since the hillslope gradient and morphology tend to vary across a landscape, it is expected that the erosion and sediment delivery would also be non-uniform. In this study, we explore the probability of the flux at a particular point in the catchment reaching the river mouth using connectivity and the Revised Universal Soil Loss Equation (RUSLE) in the Pranmati river catchment (a small 3rd order Himalayan river catchment within the Ganga River system). Methodology involves characterising the hillslopes of Pranmati river catchment centered on land use and land cover units. Using RUSLE, the sediment yielding capacity of various land cover units are estimated based on which potential source areas are marked. The sediment connectivity within the basin is also calculated by generating a sediment connectivity map of the area using method given by Borcelli et al. (2008). The catchment is categorized into four classes – (A) Highly connected zones with high sediment yielding capacity (B) highly connected zones but low yielding capacity (C) poorly connected zones but high yielding capacity (D) poorly connected zones and low yielding capacity. The area is then mapped on the basis of the defined classes and potential areas of erosion and storage are identified. Our results show that about 62% of the catchment area has low connectivity implying sediment flux generated in these zones have a low probability of leaving the catchment. Only 11% of the catchment area has sediment yield greater than the mean yield per hectare. The sediment generated from this small area of the catchment contributes 93% of the total sediment production of the catchment. References Borselli, L., Cassi, P., & Torri, D. (2008). Prolegomena to sediment and flow connectivity in the landscape: a GIS and field numerical assessment. Catena, 75(3), 268-277. Gilbert, G. K. (1877). Geology of the Henry mountains (pp. i-160). Government Printing Office.

Enhanced water management systems depend on accurate estimation of hydraulic properties of subsurface formations. This is while hydraulic conductivity of geologic formations could vary significantly. Herein, we studied an intensively managed area located in the Upper Sangamon Watershed in Central Illinois, U.S.A., and generated 2D maps of hydraulic conductivity over a large-scale region with quantified uncertainties in different depth layers. In doing so, we made use of low cost, small-scale measurements obtained from the Electrical Earth Resistivity together with more accurate, more expensive pumping tests in a calibration framework based on Kriging. We offered a cost-effective approach to reliably characterize the hydraulic conductivity properties in under-sampled sites and can be particularly used in obtaining large-scale parameter maps for a region using small-scale measurements in an efficient way. This work also includes optimal sensor placement, where the best locations for future data collection are selected by considering the current confidence levels estimated by the Kriging model, which is related to the expected value of information from future sensor data. Our approach is based on the Bayesian experimental design, which selects the best locations, out of a set of candidate locations, based on the value of information that each location is expected to offer.

Turbulence generated by aquatic vegetation plays a vital role in the interfacial transfer process at the air-water interface and sediment-water interface (AWI and SWI), impacting the dissolved oxygen (DO) level, a key indicator of water quality for aquatic ecosystems. We investigated the influence of vegetation, under different submergence ratios and plant densities, on the interfacial gas transfer mechanisms. We conducted laboratory experiments in a unidirectional recirculating flume with simulated rigid vegetation on a sediment bed. Two-dimensional planar Particle Image Velocimetry (2D-PIV) was used to characterize the mean flow field and turbulent quantities. Gas transfer rates at the AWI were determined by monitoring the DO concentration during the re-aeration process in water. SWI interfacial transfer fluxes were estimated by measuring the DO concentration difference between the near-surface and near-bed values. Compared to previous observations on a smooth bed without sediment, the presence of sediment enhances the bottom roughness, which generates stronger bed-shear turbulence. The experimental result shows that turbulence generated from the bed does not affect the surface transfer process directly. However, the near-bed suspended sediment provides a negative buoyancy term that reduces the transfer efficiency according to the predictions by a modified Surface Renewal model for vegetated flows. The measured interfacial transfer fluxes across the SWI show a clear dependence on the within-canopy flow velocity, indicating that bed shear turbulence and within-canopy turbulence are critical indicators of transfer efficiency at SWI in vegetated flows. A new Reynolds number dependence model using near-bed turbulent kinetic energy as an indicator is proposed to provide a universal prediction for the interfacial flux across the SWI in flows with aquatic vegetation. Our study provides critical insight for future studies on water quality management and ecosystem restoration in natural water environments such as lakes, rivers, and wetlands.

Stream water temperature is considered a “master variable” in environmental processes and human activities. Existing process-based models have difficulties with defining true equation parameters, and sometimes simplifications like assuming constant values influence the accuracy of results. Machine learning models are a highly successful tool for simulating stream temperature, but it is challenging to learn about processes and dynamics from their success. Here we integrate process-based modeling (SNTEMP model) and machine learning by building on a recently developed framework for parameter learning. With this framework, we used a deep neural network to map raw information (like catchment attributes and meteorological forcings) to parameters, and then inspected and fed the results into SNTEMP equations which we implemented in a deep learning platform. We trained the deep neural network across many basins in the conterminous United States in order to maximize the capturing of physical relationships and avoid overfitting. The presented framework has the ability of providing dynamic parameters based on the response of basins to meteorological conditions. The goal of this framework is to minimize the differences between stream temperature observations and SNTEMP outputs in the new platform. Parameter learning allows us to learn model parameters on large scales, providing benefits in efficiency, performance, and generalizability through applying global constraints. This method has also been shown to provide more physically-sensible parameters due to applying a global constraint. This model improves our understanding of how to parameterize the physical processes related to water temperature.