Image source: Visual Data Analytics, LLC

(1)

Richard Koehler, PhD, PH

Visual Data Analytics, LLC

Using a Temporal Information System

for Visualization and Analysis of

Hydrologic Time

Hydrologic Time--Series Data

Series Data

Application and Case Study Results

Using a Temporal Information System

for Visualization and Analysis of

Hydrologic Time

Hydrologic Time--Series Data

Series Data

Application and Case Study Results

Image source: www.rivers.gov

(2)

Discovery consists of looking at the

same thing as everyone else and

thinking something different.

Albert Szent-Györgyi

Nobel Laureate

(3)

GIS and time-series

Landscape + line graphs

Timescape!

Temporal maps

(4)

Traditional hydrograph

*

Glen Canyon Dam

Colorado River at Lees Ferry, AZ

(5)

Raster hydrograph

*

* Glen Canyon Dam

Colorado River at Lees Ferry, AZ

(6)

Raster hydrograph

*

* Glen Canyon Dam

1. Snowmelt runoff 2. Drought 3. Storm 4. Vegetation signal 5. Tunnels closed 6. El Nino 7. Artificial flood 8. Sundays 9. Christmas 10. Monthly change

Temporal patterns

Colorado River at Lees Ferry, AZ

92 years of daily data (33,700 values)

1 2 3 4 5 6 7 8 9 10

(7)

(8)

Application: QA/QC

Missing

Fort Peck computed daily inflow (~72 yrs, 26,300 values)

CAN MT ID _WY Switch to DST, systematic error Switch to DST, systematic error Switch to DST, systematic error

(9)

Application: Multi-site comparison

Drier Wetter Drought index Oregon

(10)

Month Year

Inter-comparisons

Temporal persistence Spatial extent Value color

1 2 3 4 5 6 7 8 9

(11)

Red River of the North at Fargo, ND

Symbol size ∝ flow

Flow (cfs) Year

Max = 29,300 2009

Min = 323 1934

(12)

Colorado River Water Availability Study (CRWAS)

How can climate simulations for

water resources be re-purposed?

Case study 1

Topics

Hydrometeorology Climate change Water supply Consumptive use Decision support Reservoir management Instream flows Data visualization Elements a. Sites = 845 Diversions, Reservoirs, Stream gages, ISF reaches, Natural Flow Nodes,…

b. Parameters = 26

Demand, CU, Loss, Flow,…

c. Climate scenarios = 11

(13)

Current data display

a. Colorado River nr

CO-UT state line

b. Upstream Inflow

c. All climate scenarios

Elements

a. Sites

1

b. Parameters

1

(14)

Proposed data display

CO-UT state line

b. Upstream Inflow

c.

Historic climate scenario

Elements

a. Sites

1

b. Parameters

1

(15)

CO-UT state line

b. Upstream Inflow

c.

2070 G climate scenario

Elements

a. Sites

1

b. Parameters

1

c. Scenarios

1

(16)

CO-UT state line

b.



Upstream Inflow (new)

c. (2070 G) – (Historic)

a. Sites

1

b. Parameters

1

c. Scenarios

2

Elements

(17)

1. 2. 3. 4. 5.

a. Sites

1

2

1

multiple

b. Parameters

1

2

1

multiple

c. Scenarios

1

2

multiple

1. Temporal signature

2. Up and downstream - or - basin to basin comparison 3. Dual parameter comparison

4. Scenario difference comparison (as shown earlier) 5. More complex intercomparisons

Potential new products

Elements

(18)

Case study 2

Paralytic shellfish toxins in Puget Sound*

* Moore, S.K., et al., 2009. Recent trends in paralytic shellfish toxins in Puget Sound, relationships to climate, and capacity for prediction of toxic events. Harmful Algae 8, 463–477 doi:410.1016/ j.hal.2008.1010.1003.

CAN WA

Topics

Water quality Habitat monitoring Hydrometeorology Oceanography Public health Economics Climate change Decision support

Ecological forecasting and trends Data visualization

How can multiple environmental time-series

be integrated into a single summary plot?

(19)

Traditional plots:

Background information

Time-series datasets:

Environmental factors

1.

Sea surface temp (



C)

2. Sea surface salinity (psu)

3. Air temp (



C)

4. Precipitation (cm)

5. Streamflow (m

3

s

-1

)

6. Tidal height difference (m)

7. Upwelling (m

3

s

-1

100 m

-1

)

8. Wind speed (ms

-1

)

(20)

Threshold example

Observed streamflow

1,716 days

Criterion: Flow ≤ 350 m

3

s

-1

Met = 1, Not met = 0

Missing

“Event Windows”

(21)

Case study results

1,840 days 2,513 days 2,424 days 2,546 days 2,662 days 1,716 days 2,062 days 4,116 days

Potential Event Windows

Day of Year

Calend

ar Y

ea

r

127 days

Apply specific criterion to specific layer

If



layers = 8 for any day;

Then event window present

(22)

Benefits of temporal maps

•

Natural way to view large datasets

•

Quickly review and interpret

•

Develop new types of products

•

Cost effective and time efficient method

(23)

Questions?

Visual Data Analytics, LLC

Richard Koehler, PhD, PH

[email protected]

720-840-4237

Visit our website for more information.