This Year's Projects

Schedule

The 2018 EISI program will take place June 18 – Aug. 24, 2018

Orientation will take place during the week of June 18th and final presentations will take place August 24th. Please contact the EISI Program Manager if you have any questions about the schedule.

Research areas

Students will work as a member of a research team on one of the projects described below. In the Letter of Interest required for the EISI application, students are required to identify their preference for the projects and explain how you will contribute to and benefit from each project.

Plant-pollinator networks in a changing forest landscape

Project mentors: Julia Jones (OSU-Geosciences), Andy Muldenke (OSU), Rebecca Hutchinson (OSU - Computer Science)

Scientists have witnessed widespread declines in the European honeybee and accumulated evidence that native pollinators contribute to crop yield. Networks of pollinators and the plants they pollinate exhibit complex interactions that contribute to evolution and dispersal of both the plants and the pollinators. Few locations provide intact, undisturbed pollinator communities that allow us to study these networks of interactions.

Montane meadows of the western Cascade Range of Oregon represent largely undisturbed plant-pollinator networks. These meadows occupy only a small percent of area of the western Cascades, but they contain a very large proportion of the diversity of insects and plants in the landscape. Meadows contain diverse and dynamic networks of hundreds of pollinator species (insects, birds, and other organisms) who visit dozens of flowering plant species during the spectacular, but short-lived flowering period each summer. Ongoing measurements of these pollinator networks provide the opportunity to ask key questions linking ecology, mathematics, and computer science:

  1. How do the size, location, and characteristics of meadows influence the structure of their plant-pollinator networks?
    1. Many pollinator networks have the property of "nestedness" (specialist pollinators interact with generalist plants and specialist plants are pollinated by generalist pollinators). Is "nestedness" a good measure of resilience? How does the nestedness of pollinator networks vary according to meadow size and isolation?
    2. How is the complexity of pollinator networks related to species diversity and the stability and resilience of pollinator networks? Do networks that are larger (more species-rich) have more complexity in their structure? How do networks respond to the loss of a plant or pollinator species?
    3. How does the changing climate (varying amounts of winter snow, precipitation, and temperature) influence plant-pollinator networks?
    4. How do intrinsic properties of meadows (soil moisture, size, etc.) relate to characteristics and stability of the plant-pollinator networks?
    5. How do non-native plant and pollinator species influence plant-pollinator network structure?

  2. How do the properties of modeled plant-pollinator networks compare to observed networks?
    1. How imperfect are our observed networks? Site occupancy modeling provides a way of accounting for errors in field observations. Can we adapt site occupancy models to plant-pollinator data to learn about detectability of plants, pollinators, and their interactions?
    2. We can model a hypothetical "true" plant-pollinator network with specific mathematical properties of theoretical networks and compare the modeled network to observed networks.
    3. We can represent the plant-pollinator network as a system of preferences (of plants for pollinators or vice-versa), and model the network using computational tools similar to those used to recommend movies on Netflix.

Students sample plants and pollinators in montane meadows of the HJ Andrews Experimental Forest.

         

Hummingbirds are frequent pollinators of columbine in montane meadows (Source).
Butterflies and moths also are important flower visitors.

Many species of bees and flies pollinate commonly occurring flowers.

Network drivers of threshold sediment entrainment in rivers

Project mentors: Desirée Tullos (OSU-Biological and Ecological Engineering), Catalina Segura (OSU-Forest Engineering, Resources & Management), Rebecca Hutchinson (OSU - Computer Science)

The grain size of a stream bed is a key component of habitat for stream organisms, but predicting stream grain size is challenging. The fields of river engineering, restoration, and geomorphology rely on the assumption that channels exist under an equilibrium condition, whereby the median grain size on the surface of the bed are only mobilized during bankfull flow conditions. Under this assumption the shear stress at bankfull flow (t*, usually assumed to occur every 1-2  years) is equal to the critical stress required to mobilize the median grain size on the bed surface (t*c). However, recent field studies indicate that many natural and modified channels do not meet this equilibrium condition. Instead, it appears that channels experience bankfull shear stresses much greater than the critical shear stress required to mobilize its median grain size.  Recent investigations indicated that differences in sediment supply may explain this discrepancy. Many factors can alter the sediment supply including glacial and mass movement (e.g. landslides, debris flows) history.  In addition, the presence of features such as large wood and tributary junctions, may contribute to these departures from the equilibrium condition.

The broad research theme for this project is the investigation of when, where, and potentially why channels across the HJ Andrews do and do not behave in a threshold manner, recognizing that many factors contribute to channel behavior.  More specifically, field studies will investigate: 

1)   At which locations within HJA do rivers behave as threshold channels, and which channel and landscape factors correlate with bankfull to critical shear stress?

2)    What is the relationship between landform history (e.g., glaciation) and geomorphic processes across the HJA with sediment transport stage (t*/t*c).  

3)    How do the distributions of physiographic variables derived from LiDAR (e.g., slope, elevation, valley width) and field data (e.g., large wood, confluences) explain landscape variability in transport stage across the network? 

Students working on this project will conduct field work, analyze remotely-sensed geospatial data, build models of channel hydraulics, and develop statistical and computational (i.e. machine learning) models of variables driving sediment transport across river networks. 

    

 

Riparian corridor transport in a coupled surface-subsurface system 

Project mentors: Adam Ward (Indiana Univ. - Public and Environmental Affairs), Skyuler Herzog (Indiana Univ. - Public and Environmental Affairs), Rebecca Hutchinson (OSU - Computer Science)

The exchange of water, solutes, and energy between rivers and their landscapes is crucial to water quality, ecosystem health, and ecosystem services derived from the river corridor. The connectivity between surface waters, hyporheic zones, riparian zones, floodplains, hillslopes, and aquifers – collectively termed “river corridor exchange”– is essential to these ecological benefits. In this study we consider how (imperfect) field observations at one scale can be transferred in space (to different study reaches), time (to different discharge conditions), and aggregated to use reach-scale studies to make network-scale predictions. 

The project will involve many days of field work in the creeks and rivers at HJ Andrews. Students will conduct solute tracer experiments and compare them to reduced complexity models of transport. The team will leverage an existing database of hundreds of solute tracer studies from around the catchment as a basis for prediction of exchange fluxes and timescales throughout the basin. The key outcome will be to understand the critical places and times to observe the system, and to quantify the unique information that is gained by conducting an additional field experiment (i.e., the unique information content and value of additional field information).