|
|
||
 INFLOW
ACCUMULATION
EROSION
INFLOW
EROSION
TRANSPORT
OUTFLOW
ACCUMULATION
Figure 2.1.3 Zonation of reservoir sediments
2.1.5 DATA ANALYSIS
Data analysis should consist of a well-defined program of data management, verification,
preliminary assessments, and appropriate analyses. Data management includes the collection, storage,
and distribution of the data and often requires (certainly is enhanced by) the oversight of a single
individual to insure quality handling. Typically, data are initially recorded on field sheets or electronically
in database managers or data loggers. These data are then transferred to a more permanent database.
Verification at this point is best accomplished with visual review of the data in the database and the
original data record. Other useful techniques are screening programs that use minimum and maximum
values to identify outliers (usually errors during recording or faulty conversions to different units of
measurement). For many constituents, environmentally unrealistic values may be used to identify
inappropriate values. For instance, a temperature of 80o F would easily be identified as an
inappropriate value if the temperature in the database was recorded in oC and data entry failed to
convert the field measurement. Storage of the data is also important especially for distribution to users.
A format and media must be selected that is easily transferred. Proper documentation (e.g., units of
measure, site location, analytical methods, etc.) should be provided with the data or as a separate file.
Preliminary assessments include visual data displays such as scatter plots and simple statistical analyses
such as tests for normality, homogeneity, and independence. Once distributions and other assumptions
have been evaluated, other statistical techniques (nonparametric and parametric regression analyses,
principal components analysis, multivariate analysis, etc.) can be used for more rigorous analysis and
hypothesis testing.
An important point about data analysis is that final reporting of the analysis should consider
presenting the results clearly and in a format easily interpreted by scientists, managers, and others,
2.1-10
|
||
|
||
