What is Dendra?
Dendra is an online cloud service used for retrieving, storing, managing, and curating time-series environmental data from sensor networks.
Dendra facilitates the efficient research and monitoring of our environment through real-time sensor network management, data cleaning and exploration. It supports the needs of small research groups, large state agencies, and the general public.
Dendra was designed to help organizations with long-term monitoring. It enables massive scaling of sensor deployments without a corresponding massive increase in staffing.
Whether data were logged a minute ago or decades prior, if it has a timestamp, Dendra can retrieve and graph it.
Our History
Section titled “Our History”Dendra has been in operation since 2016. As of early 2026, Dendra
- Has been adopted by 12 organizations
- Manages 300 stations with roughly 8,000 datastreams
- Stores over 5 billion records
---
config:
venn:
height: 900
---
venn-beta
title What is Dendra?
set A["Curation & Quality Control"]
set B["Data Repository"]
set C["Day-to-Day System Management"]
union A,B
union B,C
union A,C
union A,B,C["Dendra"]
style A fill:yellow
style B fill:lightblue
style C fill:red
Our Motivation
Section titled “Our Motivation”Evolving Technologies
Section titled “Evolving Technologies”In the past decade, sensors and monitoring equipment have become radically cheaper, and the technical barrier to implementation has dropped. During the same period, wireless network access has gone from rare to common. This has come in the form of carrier-grade WiFi, cellular networks, satellite uplinks, and LoRaWAN networks. This combination has heralded the advent of the Internet of Things (IoT) in the commercial world (Gubbi, 2013).
There has been a parallel, albeit slower, expansion in environmental monitoring in the academic world. The deployment of sensor observatories has become practical.
Tech companies have made a lot of advancements with IoT, but real-time data are used primarily for alerts and triggering actions. The data are stored, but there is little business incentive to QA/QC the archived data. Similarly, Big Data is focused on finding a signal through the noise within an enormous mass of data. All non-signal datapoints are ignored.
There is little business incentive for companies to develop quality control systems for the stored data. As a result, there are no time-series curation systems available commercially.
Scaling Up in the Sciences
Section titled “Scaling Up in the Sciences”For the field sciences, the quality of the data is central to our work. The proper curation of time-series data is critical. Each datapoint happens in a unique moment in time and space that will never happen again.
With climate change causing rapid environmental change across landscapes, we need to know the difference between a state change in a microclimate and an instrumentation error. Capturing as much context as to why a signal looks different can mean the difference between discovering a breakthrough finding and hoarding junk data.
In the past, careful management of field instrumentation wasn’t inherently problematic. Scientists would visit sites regularly, manually download data, and take field notes. Quality control was and still often is typically done either by the researcher when they are working on a paper, or a data manager who often has never been out in the field.
However, as the number of sensors grows from tens of instruments into thousands, manually checking and correcting datasets becomes impossible without an enormous investment in manpower. By integrating the management of a sensor observatory, the acquisition systems, and the distribution system into a single service, we can radically improve the operation and quality of data coming from a sensor observatory.
Real-time data is not valuable for its immediacy, but because the retrieval of data is automated and status of the equipment is monitored.