Skip to main content
eScholarship
Open Access Publications from the University of California

Scholarly Works (11 results)

  • Article
  • Peer Reviewed
UC Berkeley Previously Published Works (2016)
    Cover page: Optimizing embedded sensor network design for catchment-scale snow-depth estimation using LiDAR and machine learning
    • Thesis
    • Peer Reviewed
    UC Berkeley Electronic Theses and Dissertations (2017)

    California relies extensively on two pieces of natural infrastructure for water storage and conveyance: the Sierra Nevada mountains and the Sacramento-San Joaquin Delta. The mountains act as a natural reservoir, storing winter precipitation throughout the year and slowly releasing it during the summer when demand is high. Snowmelt is then routed through the Sacramento-San Joaquin Delta to the state water project, which distributes water throughout California. Despite the importance of these natural environments for the state's water resources, the in-situ infrastructure for water monitoring is limited. The existing distribution of real-time snow-sensors is present only in flat, low-elevation regions, which does not adequately capture the spatial variability of snow depth in complex terrain. To address this, the state relies on monthly synoptic snow surveys, which require trained surveyors to manually measure the snow depth at locations across the Sierra Nevada. This process provides dense in-situ data but is labor intensive and temporally sparse. It also does not capture other important variables, such as soil moisture and radiative forcing, which are necessary for next-generation hydrologic modeling. In the Delta, in-situ measurements are limited to Eulerian sensors, which measure the rate of flow going past a point on shore, but are unable to resolve finer resolution flow fields, or measure properties of the water, such as temperature, turbidity, and salinity. This dissertation explores problems related to the design and optimization of wireless in-situ sensor networks for monitoring the water balance in the Sierra Nevada and flow fields in the Sacramento-San Joaquin Delta. Topics include: optimizing wireless snow-sensor placements using LIDAR data and machine learning, constructing reliable wireless mesh networks in complex terrain, estimating the spatial variability of soil moisture in montane regions from in-situ sensors, and the design and controller optimization of wireless in-situ sensors in the Sacramento-San Joaquin Delta.

      Cover page: Design and Optimization of Wireless-Sensor Networks for Real-Time Monitoring in the Sierra Nevada and Sacramento-San Joaquin Delta
      • Article
      • Peer Reviewed
      UC Berkeley Previously Published Works (2017)
      We evaluate the accuracy of a machine-learning-based path loss model trained on 42 157 324 RSSI samples collected over one year from an environmental wireless-sensor network using 2.4 GHz radios. The 2218 links in the network span a 2000 km 2 basin and are deployed in a complex environment, with large variations of terrain attributes and vegetation coverage. Four candidate machine-learning algorithms were evaluated in order to find the one with lowest error: 1) Random Forest; 2) AdaBoost; 3) Neural Networks; and 4) K -Nearest-Neighbors. Of the candidate models, Random Forest showed the lowest error. The independent variables used in the model include path distance, canopy coverage, terrain variability, and path angle. We compare the accuracy of this model to several well-known canonical (free space, plane earth) and empirical propagation models (Weissberger, ITU-R, COST235). Unlike canonical models, machine-learning algorithms are not problem-specific: they rely on an extensive dataset and a flexible model architecture to make predictions. We show how this model achieves a 37% reduction in the average prediction error compared to the canonical/empirical model with the best performance. This paper presents an in-depth discussion on the strengths and limitations of the proposed approach as well as opportunities for further research.
        Cover page: A Machine-Learning-Based Connectivity Model for Complex Terrain Large-Scale Low-Power Wireless Deployments
        • Article
        • Peer Reviewed
        UC Merced Previously Published Works (2016)
        We assessed the potential value of hydrologic forecasting improvements for a snow-dominated high-elevation hydropower system in the Sierra Nevada of California, using a hydropower optimization model. To mimic different forecasting skill levels for inflow time series, rest-of-year inflows from regression-based forecasts were blended in different proportions with representative inflows from a spatially distributed hydrologic model. The statistical approach mimics the simpler, historical forecasting approach that is still widely used. Revenue was calculated using historical electricity prices, with perfect price foresight assumed. With current infrastructure and operations, perfect hydrologic forecasts increased annual hydropower revenue by $0.14 to $1.6 million, with lower values in dry years and higher values in wet years, or about $0.8 million (1.2%) on average, representing overall willingness-to-pay for perfect information. A second sensitivity analysis found a wider range of annual revenue gain or loss using different skill levels in snow measurement in the regression-based forecast, mimicking expected declines in skill as the climate warms and historical snow measurements no longer represent current conditions. The value of perfect forecasts was insensitive to storage capacity for small and large reservoirs, relative to average inflow, and modestly sensitive to storage capacity with medium (current) reservoir storage. The value of forecasts was highly sensitive to powerhouse capacity, particularly for the range of capacities in the northern Sierra Nevada. The approach can be extended to multireservoir, multipurpose systems to help guide investments in forecasting.
          Cover page: Valuing year‐to‐go hydrologic forecast improvements for a peaking hydropower system in the Sierra Nevada
          • Article
          • Peer Reviewed
          UC Merced Previously Published Works (2018)
          We developed an approach to estimate snow water equivalent (SWE) through interpolation of spatially representative point measurements using a k-nearest neighbors (k-NN) algorithm and historical spatial SWE data. It accurately reproduced measured SWE, using different data sources for training and evaluation. In the central-Sierra American River basin, we used a k-NN algorithm to interpolate data from continuous snow-depth measurements in 10 sensor clusters by fusing them with 14 years of daily 500-m resolution SWE-reconstruction maps. Accurate SWE estimation over the melt season shows the potential for providing daily, near real-time distributed snowmelt estimates. Further south, in the Merced-Tuolumne basins, we evaluated the potential of k-NN approach to improve real-time SWE estimates. Lacking dense ground-measurement networks, we simulated k-NN interpolation of sensor data using selected pixels of a bi-weekly Lidar-derived snow water equivalent product. k-NN extrapolations underestimate the Lidar-derived SWE, with a maximum bias of −10 cm at elevations below 3000 m and +15 cm above 3000 m. This bias was reduced by using a Gaussian-process regression model to spatially distribute residuals. Using as few as 10 scenes of Lidar-derived SWE from 2014 as training data in the k-NN to estimate the 2016 spatial SWE, both RMSEs and MAEs were reduced from around 20–25 cm to 10–15 cm comparing to using SWE reconstructions as training data. We found that the spatial accuracy of the historical data is more important for learning the spatial distribution of SWE than the number of historical scenes available. Blending continuous spatially representative ground-based sensors with a historical library of SWE reconstructions over the same basin can provide real-time spatial SWE maps that accurately represents Lidar-measured snow depth; and the estimates can be improved by using historical Lidar scans instead of SWE reconstructions.
            Cover page: Spatial snow water equivalent estimation for mountainous areas using wireless-sensor networks and remote-sensing products
            • Article
            • Peer Reviewed
            UC Berkeley Previously Published Works (2020)
            Radiation air dose rates near the Fukushima Daiichi Nuclear Power Plant (FDNPP) have been steadily decreasing over the past eight years since the release of radioactive elements in March 2011. Currently, the radiation monitoring program is expected to transition to long-term monitoring after most of the remediation activities are completed. The main long-term monitoring objectives are to (1) confirm the continuing reduction of contaminant and hazard levels, (2) provide assurance for the public, (3) accumulate the basic datasets for scientific knowledge and future preparation, and (4) detect changes or anomalies in contaminant mobility (if they occur), or any unexpected processes or events. In this work, we have developed a methodology for optimizing the monitoring locations of radiation air dose-rate monitoring. Our approach consists of three steps in order to determine monitoring locations in a systematic manner: (1) prioritizing the critical locations, such as schools or regulatory requirement locations, (2) diversifying locations that cover the key environmental controls that are known to influence contaminant mobility and distributions, and (3) capturing the heterogeneity of radiation air-dose rates across the domain. For the second step, we use a Gaussian mixture model to identify the representative locations among multiple environmental variables, such as elevation and land-cover types. For the third step, we use a Gaussian process model to capture and estimate the heterogeneity of air-dose rates across the domain. Employing an integrated dose-rate map derived from Bayesian geostatistical methods as a reference map, we distribute the monitoring locations in such a way as to capture the heterogeneity of the reference map. Our results have shown that this approach allows us to select monitoring locations in a systematic manner such that the heterogeneity of air dose rates is captured by the minimal number of monitoring locations.
              Cover page: Optimizing long-term monitoring of radiation air-dose rates after the Fukushima Daiichi Nuclear Power Plant
              • Article
              • Peer Reviewed
              UC Berkeley Previously Published Works (2021)
              Common methods for spatial distribution, such as hydrologic response units, are subjective, time-consuming, and fail to capture the full range of basin attributes. Recent advances in statistical-learning techniques allow for new approaches to this problem. We propose the use of Gaussian Mixture Models (GMMs) for spatial distribution of hydrologic models. GMMs objectively select the set of modeling locations that best represent the distribution of watershed features relevant to the hydrologic cycle. We demonstrate this method in two hydrologically distinct headwater catchments of the Sierra Nevada and show that it meets or exceeds the performance of traditionally distributed models for multiple metrics across the water balance at a fraction of the time cost. Finally, we use univariate GMMs to identify the most-important drivers of hydrologic processes in a basin. The GMM method allows for more robust, objective, and repeatable models, which are critical for advancing hydrologic research and operational decision making.
                Cover page: Optimizing spatial distribution of watershed-scale hydrologic models using Gaussian Mixture Models
                • Article
                • Peer Reviewed
                UC Berkeley Previously Published Works (2016)
                Next-generation hydrologic science and monitoring requires real-time, spatially distributed measurements of key variables including: soil moisture, air/soil temperature, snow depth, and air relative humidity. The SierraNet project provides these measurements by deploying low-power mesh networks across the California Sierra Nevada. This demo presents a replica of the end-to-end SierraNet monitoring system deployed in the Southern Sierra. This system is a highly reliable, low-power turn-key solution for environmental monitoring.
                  Cover page: SierraNet
                  • Article
                  • Peer Reviewed
                  UC Berkeley Previously Published Works (2018)
                  Using 6 yr (Water Year [WY] 2009–WY 2014) of hourly in situ measurements from a spatially distributed water-balance cluster, we quantified the long-term accuracy of an algorithm used to predict spatial patterns of depth-integrated soil-water storage within the rain–snow transition zone of the southern Sierra Nevada. The algorithm—the multivariate, non-parametric regression-tree estimator Random Forest—was used to predict soil-water storage based on a combination of attributes at each instrument cluster (soil texture, topographic wetness index, elevation, northness, and canopy cover). Out-of-bag R2 (similar to cross-validation for Random Forest) was used to quantify the accuracy of the estimator for unobserved data. Accuracy was consistently high during the wet-up, snow-cover, and early recession periods of average and wet years. The accuracy declined at the end of a 3-yr dry period, and the relative rank of the independent variables in the model shifted. Soil texture was the highest-ranked independent variable across all years, followed by elevation and northness. Topographic wetness increased in importance during dry periods. Northness exhibited high importance during the wet-up and early recession periods of most water years. During dry years, the importance of elevation declined. In dry years, notable differences in soil-water storage at each depth include lower-than-average storage in the deeper regolith at the beginning of the water year and lower storage in near-surface layers during the winter resulting from transient snow cover.
                    Cover page: Long‐Term Variability of Soil Moisture in the Southern Sierra: Measurement and Prediction
                    • Article
                    • Peer Reviewed
                    UC Berkeley Previously Published Works (2016)
                    This paper introduces SOL, which is both an efficient data representation for sensor measurements and network statistics, and a complete low-power wireless sensor management system that builds around it. A SOL system consists of multiple low-power wireless mesh networks in which motes connected to sensors and actuators send data to a single server. It offers multi-tier data replication and the associated data recovery. SOL is used in 3 pilot deployments for micro-climate monitoring, building automation and agriculture applications. In conjunction with Metronome Systems' NeoMote and Manager devices, SOL is a turn-key ready-to-deploy end-to-end system.
                      Cover page: SOL: An End-to-End Solution for Real-World Remote Monitoring Systems
                      Top