WMOD00593

• Contract/Permit #: 14-06-D-7052
• Title: Ecological Impacts of Snowpack Augmentation in the San Juan Mountains, Colorado
• Date: 3/1/1976
• State: CO
• Weather Modification - Doc Type: Report

as suggested by Morel-Seytqux and Saheli (1973) for
defining annual hydrologic Iresponses to snowpack
augmentation. If winter cloud seeding were to be
conducted on a longer time ,scale than the life of
the Upper Colorado Pilot Ptoject, a time trend
analysis might seem more attractive than it did at
the start of SJEP, but there would remain the
problem of identifying cl04d seeding as the cause of
observed changes. This is ,a more serious problem
in ecological work than it ,is in a hydrologic con-
text where the link between an augmented snowpack
and stream discharge (the measured response) is
relatively more direct.

The San Juan Ecology Project research has provided a
beginning to long-term eco~ogic monitoring of a
classical nature in the surveys it produced. A
continuation of this kind Qf work may be required
if operational cloud seedi~g were applied to the
San Juan Mountains, and is 'required by the National
Environmental Protection Act (NEPA).

The decision to reject classical experimental designs
did not preclude their application in specific
experimental situations within parts of SJEP. For
example, monitoring-type st~dies have been under-
taken in the work on silver, disposition, and on tree
growth in the dendrochronology, tree biomass, and
forest phytosociology proje~ts. In yet other projects,
a classical approach has be~n usefully applied in
studying processes on a smatl spatial scale. At
this level, it becomes an e~perimental procedure
with fewer problems of control than when it is
applied to entire stream ca~chments.

Process Studies

Most of the work undertaken' in SJEP has been carried
out with the objective of eiucidating the mechanisms
by which snowpack augmentat~on may influence the
ecosystem. More realistica~ly, it has been applied
to parts of the system judg~d to be susceptible or
important. Experimental and field methods frequently
include some of those alreaay described as "Classical
Approaches," e.g. in the ma~ipulation of the snow
cover on an experimental plbt to simulate a greatly
increased snowpack, and in the comparison of
responses on this plot to those on an adjacent
control plot. However, this experimentation is done
on a very small scale rather than a catchment scale.
This allows easier control ~f extraneous factors as
well as identification of the effects of the main
factors, i.e. winter snowp~ck or a seeding agent.

Much of the work reported in this volume is empirical
and corresponds to a simple; black box model. Thus,
it tends to state relations~ipsbetween the snowpack
and ecological responses which are defined statis-
tically by regression-type analyses. These are not
statements about mechanisms, except insofar as the
experience and intuition of the investigator allow
that interpretation. Such models do, however, allow
easy prediction of effects ~y changing the input to
the black box to account for the change in snowpack
conditions produced by cloud seeding. Obviously,
it would be better to open the black box, at least
partially, and identify theltrue mechanisms whereby
a change in the snowpack makes an effect, e.g.
through intermediate conditions like a change in soil
temperature, moisture statu$, or growing season
length. SJEP has made progress in this direction
(Figure 1, Chapter II) which allows predictions to
be better evaluated; it provides models with a higher
level of generality that may be capable of extra-
polation to other mountain areas.

..

In general, we feel that a "process study" approach
to evaluating the ecological effects of cloud seed-
ing is a most useful one. It can be applied in
situations of "unmodified" snowpack and, therefore,
requires neither actual cloud seeding nor prior
evaluation of cloud seeding effectiveness. In many
cases, an assumption that sampling in space is
equivalent to sampling in time is implicit in this
approach. Thus, workers in SJEP have often assumed
that the effects of increased snowfall on a site
will be equivalent to a shift along a snowpack
gradient defined by present conditions on many sites.
Such an ergodic assumption may be advantageous. It
allows predictions for conditions following the
equilibration of the ecosystem to an impact, at least
to the level that the system is now adapted to
the "natural" snow cover.

In a process study approach, it is only necessary to
assume the effectiveness of snowpack augmentation
at the stage of impact evaluation or prediction.
When a knowledge of processes is applied in this way,
a more detailed knowledge of snow conditions in the
mountains than was available to the SJEP is needed.
This deficiency has hindered impact evaluation through
much of the project.

The study of the mechanisms whereby an augmented
snowpack might have ecological significance also
allows efficient use of existing ecological knowledge.
Such knowledge may not be immediately applicable
to the question of evaluating an augmented snowpack
but can often be linked intuitively to such an
environmental change. For example, existing knowledge
of the effects of lower soil temperatures and
abbreviated growing season on plant growth can be
applied to evaluating the impact of increased
snowfall, irrespective of the source of information.

Finally, the search for cause-and-effect relation-
ships, even if they are not simple and direct ones,
is eminently suitable to a scientific, rationalist
viewpoint. This means that the statistical pro-
cedures used in hypothesis testing should be
available to workers using a process study approach.

SimulaOion Modelling

Experimentation with computer simulations was not
seriously considered as a general strategy for the
SJEP at its inception. At that time, the expense
of modelling, the undeveloped nature of available
models and the lack of professional expertise were
sufficient reasons for this decision. Since then,
significant progress in the modelling of ecosystem
dynamics has been made, especially under the Analysis
of Ecosystems Program of The International Biology
Programme. This progress has been paralleled in
hydrologic modelling, even to predicting the hydro-
logic effects of snowpack augmentation (Leaf 1975).

The development of system simulations in SJEP was
impeded by another early decision: that a full
system study could not be funded. With only parts
of any ecosystem under study, it became difficult
to attempt integrated simulation at a,later stage.

This does not mean that simulations have not been
used in SJEP, although those used have been loosely
structured "conceptual" simulations rather than
rigorous mathematical models. The use of mathe-
matical simulations has been restricted to work
within individual projects. Were SJEP to be
initiated today, our approach to simulation would
probably be different, if only to take advantage of

29

l'