This website uses the CSS Grid layout, which your browser does not support.

waders

Problem Framing

What kinds of data models can be developed to reduce the effort and risks involved in siting climate sensor towers?

Background

In climate science, flux sensor towers involve many pieces of equipment including observational instruments, reference instruments, data loggers, power sources, and others. For researchers, determining the optimal placement of each item requires collecting and interpreting data from many sources. This process can be arduous and dangerous. The DOE-funded AmeriFlux project gathers and shares long-term carbon, water and energy flux measurements for climate science from over 500 sensor towers throughout the Americas.

Other factors that could skew data include exposure to decomposing biomass, eddy shadows, precipitation, wildlife movement, instrument fatigue, etc. Eddy covariance is used to calculate fluxes from varied sources.

A wind rose is a graphical representation of the wind patterns for a geographical location. Decisions about instrument and tower siting can require periodic reevaluation in the face of disturbances such as wildfires, crop harvesting, and species migration. Data visualization coupled with UAV imagery can reduce effort in the evaluation of potential sites. In addition, data processing and interpretation can be enhanced by triangulation with visualized wind data. [1]

Wind Rose Sketches

sketches

My Roles

  • UX Researcher
  • UX Designer
  • Feature Lead

Methodologies

  • Stakeholder interviews
  • User interviews
  • Group workshops
  • Usability studies

Process

  • Generative user interviews
  • In situ task participation
  • Evaluative usability studies

Artifacts

  • Pattern library
  • Lightweight prototypes
  • Findings and recommendations

Overview

Generative user interviews were conducted to establish user needs. I performed routine sensor correlation tasks such as measuring the distance between sensors by wading deep into the slough.

Next, data sets and models were evaluated for relevance, availability and feasibility for development of visualization. Sonic anemometer data was selected as a viable source for wind rose visualizations. Sample wind roses were constructed, and presented to users during usability studies. Designs were modified, and wind roses generated for each tower site with relevant data.

Findings

  • Data Sets: Widespread calculation of flux footprint models is not feasible at this time. Lidar data¬† from drones shows promise, but drones and qualified pilots are not available for most tower sites, especially in Latin America. Access to MODIS images is hampered by API issues.
  • Promising Data: Data from sonic anemometers is available at most tower sites. Consider using that data to display wind roses for each tower site.
  • Color Palette: Variation in the color palette of satellite imagery is a challenge to wind rose legibility. Recommend that wind roses use an "unnatural" palette of bright primary and secondary colors.
  • Scale Variation: Variation in the wind speed values across tower sites is a challenge to consistency of wind rose scales. Recommend that a single scale be established for each tower site, which will allow easy seasonal comparison for each site while presenting sufficient visual precision for use.

Wind Roses

wind roses

Outcomes

  • Nine wind roses for each tower site
  • Reduced effort in siting equipment

Reflection

  • Images for Papers: A frequent need in climate research is images for presentation and publication. Images need to be easily downloadable and embeddable in common presentation formats such as PowerPoint and Google Presentations.
  • Followup: Primary Findings


UAV and Satellite Data: UAV imagery[2] can greatly reduce the effort required to select sites for instruments and towers, which can be over 30 meters tall, to be able to observe the surrounding vegetative canopy. Lidar imagery and satellite data such as this MODIS image from another site also inform siting efforts.

aerial view of trees in forest Normalized Difference Vegetation Index as map

Lidar Data: Lidar data[3] can be used to generate 3D models of canopy and terrain, saving time and effort in siting decisions.

lidar data of canopy plot from lidar data

Site and tower variability: Tower sites range from the arctic such as Poker Flat in Alaska[4] to the tropics of Guanica Forest in Puerto Rico[5]. Siting these towers and then their instruments can be challenging in even the best local weather of the year.

aerial view of trees in forest Normalized Difference Vegetation Index as map

Why this matters: Data collected is part of research that led to the identification of the source of climate change: anthropogenic greenhouse gases. Based on precise field measurements from varied locations, this team concluded that radiative forcing is directly attributable to the increase of atmospheric CO₂. This confirms predictions of the greenhouse effect due to these emissions, which are affecting the surface energy balance. See Observational Determination of Surface Radiative Forcing by CO₂ from 2000 to 2010 Nature, March 2015.

Bits

Image credits:
[1] Jonathan Thom
[2] Jonathan Dandois
[3] Gil Bohrer and Tim Morin
[4] AmeriFlux
[5] NEON Program and Battelle