Angelo Coast Range Reserve

Environmental Sensor Observatory

Collin BodeProject Manager,

Desktop Watersheds Integrative Program,

National Center for Earth-surface Dynamics

& Specialist, Integrative Biology, UC Berkeley

Angelo Coast Range Reserve

• Banff

• NCED primary field site• UC Natural Reserve System• Rapidly eroding upland watershed• South Fork Eel River• Vegetation:

• Douglas Fir, Redwood• Oak, Madrone• Chaparral

Environmental Observatory Research Value

“Mapping, Tracing, and Wireless Sensor technologies may do for ecology in the 21st century what DNA sequencing did for genetics in the 20th century.” – Beth Burnside, molecular biologist

and Vice Chancellor for Research at UCB

Spatial extents & frequency from 52 Articles from Ecology Journal 2003 -2004 (Porter, BioScience, 2005).

Definition: An environmental sensor observatory is a distributed array of sensors connected by a network that is aggregated into a single dataset.

It becomes an observatory when the combination of sensors provides information not possible to achieve by individual sensors.

Infrastructure & Workflow

NetworkingSensors Informatics

Sensor Deployment

Sensors

Sensors: Keck HydroWatch Project

PI: Inez Fung, Atmospheric Scientist, Climate modeler

Co-PI’s:– Bill Dietrich, Geomorphologist – Todd Dawson, Tree physiology– David Culler, Computer Science

• Motes, TinyOS, CENS

Purpose: Trace and understand water movement between atmosphere to subsurface mediated by vegetation.

Field Site: Rivendell – 130 x 30 meter north facing micro-watershed– Dominated by Douglas Fir– Elder Creek at base

Angelo Coast Range Reserve

Rivendell: Keck HydroWatch Project Study Site

Sensors: Keck HydroWatch, Rivendell Study Site

Sensors: Keck HydroWatch, Rivendell Study Site

Monitoring Equipment1,264 sensors as of Sept, 201013 DataLoggers

Sensor Deployment: Rivendell & Weather Stations July, 2010

Parameter Monitored Campbell Mote Others* TotalsBarometric Pressure mb 10   10Air Temp C 30 29 32 91Logger internal Temp c 11 29   40Soil Temperature 69   69Soil moisture TDR (dialectric) 100   100Soil moisture ERP (resistance) 192   192Soil Moisture SM200 30   30Water Temp C 2 1 3Gage Height meters 11 1 12Turbidity NTU 1   1Water Samples   2 2Rainfall mm 12 1 13Hail Cumulative mm 2   2Hail Intensity mmh 2   2Rel Humidity Perc 55 29 31 110Snow Depth Meters 1   1Total Solar Radiation W/m2 4 29 30** 63PAR   29 35 64Wind Direction Degrees 12 1 13Wind Speed Avg MS 18 1 19Sap Flow velocity   154 154Battery Voltage 24 29  1 54Totals 575 174 290 1039* Sap Flow Loggers, OWL logger, HOBO pendants, Isco Samplers ** HOBO pendants offline

Sensor Deployment: Rivendell

Wireless Network Architecture

Networking

Power Tree: Ursula

Power Tree: Ingrid

PowerTree: Ilean

Tree: FlattopNo solar power

Power Tree: Brandywine:

Power Tree: Treebeard

Sensors

Networks: Rivendell Study Site

BackboneNetwork

4 Wireless networks

Network Architectures1. Backbone

1. Heirarchical tree structure

2. Eion VIP110-243. 2.4ghz, 5 watts 4. VINES protocol

2. Campbell PakBus• Manually defined

relays• 900mhz, 1watt • Pakbus protocol

3. Motes• Mesh structure • 2.4ghz, extremely low

power• Single base station to

Backbone• 802.15 protocol

Cahto PeakInternetAccess Point

RivendellKeck HydroWatchResearch Site

Backbone Network (VINES)

Mote & Campbell Network

Wifi (802.11n) Network

Wilderness LodgeBuilding

Steel ResidenceBuilding Science Center

Caching ServerAC power

Networks: Wireless Backbone

REBBackbone relay

UBERBackbone relay

BOBBackbone relay

FORBBackbone relay

Peter Steel, Reserve Steward,Setting up radio and antenna

Collin Bode, configuring at 35 meters

Networks: Wireless Backbone, REB Construction

• Redwood height: 40 meters• Solar System total weight: 80kilos• Arms: very sore

Networks: REB Construction

Informatics: Berkeley Sensor Database

Informatics

Berkeley Sensor Database: ODM

CUAHSI Observations Data Model• Hydrologic Information Systems (HIS)• Similar goals as GeoCENS in data

sharing and discovery

Berkeley Sensor Database • Ginger Ogle, Database

Programmer• Core ODM schema only• SQLserver to MySQL

Management Functions• Monitoring Collections• Stations• Datastreams

Data Quality Functions• Data Quality Levels• Data Flagging• Incident Reports

34.6 million records on Sept. 22, 2010

Berkeley Sensor Database: Researcher Use of Data

Type of Access Requirements Data Format

Web Access Web browser & valid project login

Graph & CSV download

Direct MySQL Access Knowledge of SQL SQL output

Matlab Matlab MysQL connector or Mathworks Database Toolbox

Matlab matrix format

Bulk Data Dumps Valid project login Zipped, CSV, weekly refresh

http://sensor.berkeley.edu

Berkeley Sensor Database: Data Quality Levels & Flagging

Data Flag Description

P Passed sanity check

U Unchecked value, no sanity check

NV No value recorded (NaN)

VB Value below device minimum range

VE Value exceeds device minimum range

X Invalid Data: equipment malfunction

Level Definition

0 Instrument Data

1A Instrument Data + Metadata

1B Instrument Data in Sensor Units + Metadata

2 Geophysical Units

3 Geophysical Units + Space/Time Uniformly Scaled

4 Derived Data, from Multiple Variables or from a Model

NASA/EOS Data Quality Levels Data Flags

Berkeley Sensor Database: Incident Reports

Incident ID 1

Title: Bear damageStation Name(s): Riv Level 1_1, Riv Level 1_2,

Riv Level 2_1, Riv Level 2_2,Riv Level 3_1, Riv Level 3_2, Riv Level 4_1

Datastream(s): resistance probes, rain collectorwells

Data Flagged: ERPs: 290,304 values Well 6: 3,024 valuesWell 7: 3,024 values

StartTime: 2009-09-24 (+/- 1 months)EndTime: 2009-10-24 (+/- 1 days)Reported By: Bill Dietrich (2009-10-26 00:00:00)

Description: Daniella, Jasper and I visited the site this past weekend. Unfortunately a bear did some damage recently. It systematically knocked over and chewed on everyone of the plastic bag covered resistance probes. They were each snapped at the soil boundary. I haven't had a chance to check to see when this might have happened according to the online record. The bear also knocked over the rainfall collector (James says he has found teeth marks in it before). It broke the extended plastic tube on well 7 and the wooden stake holding up the plastic tarp. and it knocked off the cap of well 6. We didn't detect other damage-- but there could be some. I encourage you all to check your favorite on line device. We put the cap back on 6, propped up well 7 tube and put back a stake. and reset up the rain collector.