Abstract
Understanding the nutrient source areas in the forested ecosystems is critical for managingwatersheds and protecting downstream water resources. This requires long-term
monitoring efforts along with modeling efforts to develop a process-based understanding
of the system under different scenarios. Few contemporary models cater to the evaluation
of the impacts of various land management strategies and climate scenarios on the water
quantity and quality dynamics. Even fewer process-based models serve the outputs in a
functional and intuitive format such that it meets the needs of the watershed managers
who would want to quickly assess multiple watersheds and management scenarios.
This study: utilizes long term water quality monitoring data to assess the effects of
commercial forest management operations on stream water quality; uses laboratory experiments
to shed light on the fate and transport of phosphorus in the forest-meadow systems;
employs a process-based model to simulate phosphorus transport from a forested watershed
and tests its accuracy to model the phosphorus losses; and develops an interactive
decision support tool that translates multi-scenario, multi-watershed simulated data from
two of the widely used models in the US into information useful for watershed managers.
The first study assesses the effect of contemporary forest management activities, including
clear-cutting and thinning, on water yield and stream nitrogen and phosphorus (P)
dynamics using a quarter-century-long (1992–2016) monitoring data from a paired and
nested watershed in the interior Pacific Northwest, US. The study showed that the contemporary
forest management activities increased stream nitrate + nitrite (NO3 + NO2)
concentrations and loads following timber harvest activities, but these effects are attenuated
due to downstream uptake processes. Interestingly, the NO3 + NO2 concentration,
streamflow, and loads of NO3 + NO2 and orthophosphate (OP) from the undisturbed
control watershed also increased. However, these increases were relatively smaller than
the harvested watersheds and likely driven by climate variability or subtle forest succession changes. Furthermore, relative to post-wildfire impacts, these nutrient increases from harvested watersheds are small (∼5 to 20-fold compared to pre-disturbance) and short-lived (∼5 years).
The second study demonstrated that the vertical phosphorus transport through the
subsoil in the forest-meadow system soils does occur and it can be significant. However,
this pathway is often presumed to be a relatively minor source of phosphorus compared
to phosphorus transport through surface runoff and soil erosion. This study showed that
phosphorus leaching in the forest-meadow systems of Lake Tahoe Basin does in fact occur
and it occurs primarily in organic form. When enriched P source is present leaching
from granitic sites is larger than that from andesitic sites and granitic meadows leach the
largest amounts of phosphorus. The greatest risk of phosphorus leaching, translocation,
and potential loss via subsurface pathways occurs in granitic soils with enriched phosphorus
sources. Saturation excess runoff is an important pathway for phosphorus loss from
meadow systems as demonstrated by the losses from the exfiltration pathway.
The third study represented one of the first times the WEPP model with water quality
algorithms (WEPP-WQ) was tested on forested watersheds. The study demonstrated
that the seasonal phosphorus transport from upland sources can be simulated using a
process-based watershed model. This study found that the P sorption parameter (PSP),
P soil partitioning parameter (PHOSKD), initial labile phosphorus pool in topsoil layer
(LabileP), and P uptake distribution parameter (UPB) are some of the relatively important
and sensitive parameters for simulating phosphorus loss using the WEPP-WQ
model. The relative differences between calibrated values obtained for these parameters
in watersheds with differing soil type are well supported by the findings of isotherm experiments
in this chapter. Watershed scale modeling study also showed adequate capabilities
of the existing phosphorus routines in WEPP to simulate soluble phosphorus losses from
the watersheds. While WEPP-WQ does not account for the P contributions associated with the channel processes, the seasonality and relative trends of particulate phosphorus
were correctly predicted. A simple analysis of TP load using a fixed P concentration
associated with detached channel sediments, suggested that the absolute magnitude of
predicted particulate phosphorus from upland sources is underpredicted by the model.
This underprediction may be due to the assumed relative distribution of P in active and
stable pools that may not be appropriate for forested soils or due to the underestimation
of P enrichment ratio or a combination of both. Further investigation of model structure
is needed to identify appropriate soil P pool initialization. Significant development and
testing are needed for WEPP-WQ to be fully ready for use. Overall, this study shows that
WEPP-WQ, with its current dissolved phosphorus routines, can be an effective, processbased,
and yet parsimonious edge-of-the-hillslope effects tool for informing land and water
management decisions. Improving the particulate P predictions and making WEPP-WQ
a complete water quality prediction tool requires further developments and testing.
In the fourth and final study, a prioritization, interactive visualization, and analysis
tool (Pi-VAT) was developed to assist watershed managers with synthesizing multiscenario,
multi-watershed outputs from process-based geospatial models. Pi-VAT was
applied to output from multiple watersheds and for multiple management scenarios and
treatments from two geospatial models for watershed management: Water Erosion Prediction
Project (WEPP) and Soil & Water Assessment Tool (SWAT). This study demonstrated
the utility of Pi-VAT to examine simulated hydrologic, sediment, and water quality
response at the hillslope/hydrologic response unit (HRU) scale. In a matter of minutes,
Pi-VAT can synthesize overwhelming amounts of output from process-based models into
information useful for land and water resources managers.