Transcript power point - pptx format

Presented at the Information Management Workshop for Forest Dynamic
Plot Database 2009
Nantou, TAIWAN
June 15th, 2009
The CTFS database workshop II
Smithsonian Tropical Research Institute (STRI) Panamá
September 29 to October 6, 2008.
Participating Nations:
Brasil - 1 plot
Canada – 1 plot
Colombia – 2 plots
Ecuador – 1 plot
DR of Congo – 1 plot
US North America
Temperate:
Wisconsin – 1 plot
Maryland – 1 plot
Hawaii – 2 plots
Puerto Rico – 1 plot
Nantou, TAIWAN
Monday, June 15, 2009








A 16 ha plot
500 m N-S and 320 m E-W
16 X 25 grid
400 20 x 20 m quadrats
16 5 x 5 m sub-quadrats
per quadrat
Plant census every 5 years
4 censuses already
Measurements include:
location, size, point of
measurement, mortality,
damage
LUQ CTFS Plot data : location, dbh, mortality code, point of
measurement


Directed Neotropical and African plot sites
Present and implement the newly designed database
system
◦ Load data from each (completed) plot into the database
system
◦ Work with data reports created by the database system
◦ Work with the database editor for minor changes to the
data
◦ Demonstrate and examine the data entry program
◦ Make concrete plans for how the databases will be
distributed:
 web applications;
 sharing level up to ea. scientist
Nantou, TAIWAN
Monday, June 15, 2009

Store and manage:
◦ enormous amount of plot data
◦ store annual changes
◦ store versions: tracking the history of data; modifications
and corrections

Minimize errors:
◦ Tree measurements errors
 Mismatching errors in the database or at the field

Can be customized to add site specific data
Nantou, TAIWAN
Monday, June 15, 2009

Used of R to check
species, quadrants, codes,
match tags, etc
- LUQ spent 2-3 days
filtering data for the 3
censuses
R
Mayor Problem: tagging

1.
2.
Mayor problems
X,Y coordinate definition-local to quadrat not to plot
Dead stem vs dead tree:
CTFS database : dead trees VS LUQ : dead stems
3.
Repetitions of records having <tag, stemtag>
Solutions:
1. Reduced the X,Y coordinate in a Paradox scripts
2. Series of queries to determine that all stems were
dead,
3. Extracted records for further inspections.

Eg., Duplicate Tag;Tree is in quadrat=1013”
A false duplication error
Eg., Another stem has larger dbh? :
LUQ’s main stem was not always the one with the biggest
diameter

Are this real problems that will cause substantial error in
the analysis of these data?
Data gathering protocols?
Conceptual design?

DO WE NEED TO STANDARDIZE AT THE METHODOLOGY
AND DATA ENTRY LEVEL?

Using private online forms

Step-by-step set of forms

Using private online forms

Step-by-step set of forms
Convert species, codes, quadrant
files into CTFS’ database format
•Great quality control tool
•Easy to use, once your data is “bug” free
•PURPOSE OR CONCEPT BEHIND:
•Database designed for storage and
management of data in a standard way,
•NOT FOR SHARING WITH OTHERS
•“Forces” standardization in some ways:
•Data design: codes and measurments
definitions
•Data gathering protocols
,
My assessment of the system
• Importance of IM workshops:
• Mix IM and scientists
• Learning the system together
• Both groups discussing the scientists’
needs and use of the system
• IM and Scientists relationship
• IM reason to exist: facilitate science
• Scientists need of IM to manage large
amount of data
•When spreadsheets are just not
enough
,
Some Thoughts