Exploring the use of neural networks for biomass forecasts in the Peruvian upwelling ecosystem

A pilot study was conducted to study the ability of an artificial neural network to predict the biomass of Peruvian anchoveta Engraulis ringens, given time series of earlier biomasses, and of environmental parameters (ocenographic data and predator abundances). Acceptable predictions of three months or more appear feasible after thorough scrutiny of the input data set.

Spatial coherence and persistence in annual streamflow in the western United States [abstract]

EXTRACT (SEE PDF FOR FULL ABSTRACT): Four broad regions of the western United States within which annual streamflows exhibit strong spatial coherence are identified using principal component analysis with a varimax rotation. Geographically, the four regions encompass the Pacific Northwest, Far West-Great Basin, Central Rockies-High Plains, and Northern Great Plains. These regions are really consistent with previously documented, descriptively derived streamflow regimes as well as with general atmospheric circulation and precipitation modes of variation. Collectively, the four regional components account for nearly 63 percent of the total annual variation in western U.S. streamflow. The time history of most principal component patterns exhibit little or no persistence.

The influence of North Pacific atmospheric circulation on streamflow in the West [abstract]

EXTRACT (SEE PDF FOR FULL ABSTRACT): The annual cycle and non-seasonal variability of streamflow over a network of stations in western North America and Hawaii is studied in terms of atmospheric forcing elements. The phase lag between the annual cycle of streamflow and precipitation varies considerably over this network, as does the persistence of monthly streamflow anomalies. This lag effect appears to be largely a function of the relative amount of snow laid down in a particular basin. In addition to the rather strong annual cycle that exists in mean streamflow and its variance at most of the stations, there is also a distinct annual cycle in the autocorrelation of streamflow anomalies that is related to the interplay between the temperature and precipitation annual cycles; of particular importance is the existence of stored water in the form of a snow pack.

Periodic autoregressive moving average models for the analysis of streamflow [abstract]

EXTRACT (SEE PDF FOR FULL ABSTRACT): Streamflow values show definite seasonal patterns in their month-to-month correlation structure. The structure also seems to vary as a function of the type of stream (coastal versus mountain or humid versus arid region). The standard autoregressive moving average (ARMA) time series model is incapable of reproducing this correlation structure. ... A periodic ARMA time series model is one in which an ARMA model is fitted to each month or season but the parameters of the model are constrained to be periodic according to a Fourier series. This constraint greatly reduces the number of parameters but still leaves the flexibility for matching the seasonally varying correlograms.

A multi-basin seasonal streamflow model for the Sierra Nevada

Linear regression models are constructed to predict seasonal runoff by fitting streamflow to temperature, precipitation, and snow water content across a range of elevations. The models are quite successful in capturing the differences in discharge between different elevation watersheds and their interannual variations. This exercise thus provides insight into seasonal changes in streamflow at different elevation watersheds that might occur under a changed climate.

Streamflow and La Niña event relationships in the ENSO-streamflow core areas

EXTRACT (SEE PDF FOR FULL ABSTRACT): The high index phase of the Southern Oscillation (SO), La Niña, has not been given as much attention as its counterpart, the low index phase of the SO, El Niño. One reason may be related to the fact that many similarities exist among El Niño events but not among La Niña events. ... In this study, we focus on the influences of La Niña phenomena on streamflow anomalies ... to explore the SO-related signal over the United States.

Wave forecasts associated with hurricane landfalls in the vicinity of Cape Lookout, North Carolina

Hurricanes can cause extensive damage to the coastline and coastal communities due to wind-generated waves and storm surge. While extensive modeling efforts have been conducted regarding storm surge, there is far less information about the effects of waves on these communities and ecosystems as storms make landfall. This report describes a preliminary use of NCCOS’ WEMo (Wave Exposure Model; Fonseca and Malhotra 2010) to compute the wind wave exposure within an area of approximately 25 miles radius from Beaufort, North Carolina for estuarine waters encompassing Bogue Sound, Back Sound and Core Sound during three hurricane landfall scenarios. The wind wave heights and energy of a site was a computation based on wind speed, direction, fetch and local bathymetry. We used our local area (Beaufort, North Carolina) as a test bed for this product because it is frequently impacted by hurricanes and we had confidence in the bathymetry data. Our test bed conditions were based on two recent Hurricanes that strongly affected this area. First, we used hurricane Isabel which made landfall near Beaufort in September 2003. Two hurricane simulations were run first by passing hurricane Isabel along its actual path (east of Beaufort) and second by passing the same storm to the west of Beaufort to show the potential effect of the reversed wind field. We then simulated impacts by a hurricane (Ophelia) with a different landfall track, which occurred in September of 2005. The simulations produced a geographic description of wave heights revealing the changing wind and wave exposure of the region as a consequence of landfall location and storm intensity. This highly conservative simulation (water levels were that of low tide) revealed that many inhabited and developed shorelines would receive wind waves for prolonged periods of time at heights far above that found during even the top few percent of non-hurricane events. The simulations also provided a sense for how rapidly conditions could transition from moderate to highly threatening; wave heights were shown to far exceed normal conditions often long before the main body of the storm arrived and importantly, at many locations that could impede and endanger late-fleeing vessels seeking safe harbor. When joined with other factors, such as storm surge and event duration, we anticipate that the WEMo forecasting tool will have significant use by local emergency agencies and the public to anticipate the relative exposure of their property arising as a function of storm location and may also be used by resource managers to examine the effects of storms in a quantitative fashion on local living marine resources.

Skill of monthly forecasts during the Northern Hemisphere summer

We have performed GCM experiments using the National Meteorological Center's Medium Range Forecasting (MRF) model to study the skill of monthly forecasts during the Northern Hemisphere summer and to test the impact of sea surface temperature anomalies (SSTAs) on such forecasts. The daily skill varies a great deal. The skillful daily forecasts last from 5 to 8 days for the Southern Hemisphere and from 6 to 8 days for the Northern Hemisphere. SSTAs have positive impact on the forecasts in the tropics and surface variables, but the impact of tropical SSTAs on the extra-tropical circulation is, in general, positive but small. Overall, the initial conditions play a more important role than SSTAs in determining the forecast skill.

"Cool" vs. "warm" winter precipitation and its effect on streamflow in California

Precipitation is a difficult variable to understand and predict. In this study, monthly precipitation in California is divided into two classes according to the monthly temperature to better diagnose the atmospheric circulation that causes precipitation, and to illustrate how temperature compounds the precipitation to runoff process.

South American streamflow and the extreme phases of the Southern Oscillation

This study investigates the extent of the affect [sic] of the El Niño/Southern Oscillation on South American streamflow. The response of South American precipitation and temperature to the extreme phases of ENSO (El Niño and La Niña events) is well documented; but the response of South American hydrology has been barely studied. Such paucity of research contrasts sharply with that available on the response of North American streamflow to ENSO events.

Decadal hydroclimatic variability in the western coastal United States: temporal and spatial variations in precipitation, streamflow, and lake level

EXTRACT (SEE PDF FOR FULL ABSTRACT): We have analyzed streamflow variations recorded at 15 USGS gauging stations in California during the past 90 years or so. The anomalies (departures from the 1960-1990 mean discharge) of streamflow on annual-to-decadal time scales are strongly correlated with precipitation anomalies in each drainage basin. ... Although causes of the decadal climate (precipitation) variability are not known with certainty, the use of streamflow records may help us understand the relative strengths of moisture sources and shift of the jet stream in atmospheric circulation.

Monte Carlo simulation of streamflow of the Salt River, Arizona

EXTRACT (SEE PDF FOR FULL ABSTRACT): Evaluations of the impact of climate change (such as a greenhouse effect) upon water resources should represent both the expected change and the uncertainty in that expectation. Since water resources such as streamflow and reservoir levels depend on a variety of factors, each of which is subject to significant uncertainty, it is desirable to formulate methods of representing that uncertainty in the forcing factors and from this determine the uncertainty in the response variables of interest. We report here progress in the representation of the uncertainty in climate upon the uncertainty in the estimated hydrologic response.

Quantifying the value of improved wind energy forecasts in a pool-based electricity market

This work illustrates the influence of wind forecast errors on system costs, wind curtailment and generator dispatch in a system with high wind penetration. Realistic wind forecasts of different specified accuracy levels are created using an auto-regressive moving average model and these are then used in the creation of day-ahead unit commitment schedules. The schedules are generated for a model of the 2020 Irish electricity system with 33% wind penetration using both stochastic and deterministic approaches. Improvements in wind forecast accuracy are demonstrated to deliver: (i) clear savings in total system costs for deterministic and, to a lesser extent, stochastic scheduling; (ii) a decrease in the level of wind curtailment, with close agreement between stochastic and deterministic scheduling; and (iii) a decrease in the dispatch of open cycle gas turbine generation, evident with deterministic, and to a lesser extent, with stochastic scheduling.

Natural streamflow regime alterations: Damming of the Piave river basin (Italy)

A novel approach is proposed to estimate the natural streamflow regime of a river and to assess the extent of the alterations induced by dam operation related to anthropogenic (e.g., agricultural, hydropower) water uses in engineered river basins. The method consists in the comparison between the seasonal probability density function (pdf) of observed streamflows and the purportedly natural streamflow pdf obtained by a recently proposed and validated probabilistic model. The model employs a minimum of landscape and climate parameters and unequivocally separates the effects of anthropogenic regulations from those produced by hydroclimatic fluctuations. The approach is applied to evaluate the extent of the alterations of intra-annual streamflow variability in a highly engineered alpine catchment of north-eastern Italy, the Piave river. Streamflows observed downstream of the regulation devices in the Piave catchment are found to exhibit smaller means/modes, larger coefficients of variation, and more pronounced peaks than the flows that would be observed in the absence of anthropogenic regulation, suggesting that the anthropogenic disturbance leads to remarkable reductions of river flows, with an increase of the streamflow variability and of the frequency of preferential states far from the mean. Some structural limitations of management approaches based on minimum streamflow requirements (widely used to guide water policies) as opposed to criteria based on whole distributions are also discussed. Copyright © 2010 by the American Geophysical Union.