Geology topics

Chiu W. Yeung

Publications and source records attributed to Chiu W. Yeung.

4 recordsLinked to original sources

USGS2010

Flood-frequency estimates for streams on Kauai, Oahu, Molokai, Maui, and Hawaii, State of Hawaii

This study provides an updated analysis of the magnitude and frequency of peak stream discharges in Hawai`i. Annual peak-discharge data collected by the U.S. Geological Survey during and before water year 2008 (ending September 30, 2008) at stream-gaging stations were analyzed. The existing generalized-skew value for the State of Hawai`i was retained, although three methods were used to evaluate whether an update was needed. Regional regression equations were developed for peak discharges with 2-, 5-, 10-, 25-, 50-, 100-, and 500-year recurrence intervals for unregulated streams (those for which peak discharges are not affected to a large extent by upstream reservoirs, dams, diversions, or other structures) in areas with less than 20 percent combined medium- and high-intensity development on Kaua`i, O`ahu, Moloka`i, Maui, and Hawai`i. The generalized-least-squares (GLS) regression equations relate peak stream discharge to quantified basin characteristics (for example, drainage-basin area and mean annual rainfall) that were determined using geographic information system (GIS) methods. Each of the islands of Kaua`i,O`ahu, Moloka`i, Maui, and Hawai`i was divided into two regions, generally corresponding to a wet region and a dry region. Unique peak-discharge regression equations were developed for each region. The regression equations developed for this study have standard errors of prediction ranging from 16 to 620 percent. Standard errors of prediction are greatest for regression equations developed for leeward Moloka`i and southern Hawai`i. In general, estimated 100-year peak discharges from this study are lower than those from previous studies, which may reflect the longer periods of record used in this study. Each regression equation is valid within the range of values of the explanatory variables used to develop the equation. The regression equations were developed using peak-discharge data from streams that are mainly unregulated, and they should not be used to estimate peak discharges in regulated streams. Use of a regression equation beyond its limits will produce peak-discharge estimates with unknown error and should therefore be avoided. Improved estimates of the magnitude and frequency of peak discharges in Hawai`i will require continued operation of existing stream-gaging stations and operation of additional gaging stations for areas such as Moloka`i and Hawai`i, where limited stream-gaging data are available.

Hawaii↗

USGS2007

Water Use in Wetland Kalo Cultivation in Hawai`i

Ten cultivation areas (8 windward, 2 leeward) were selected for a kalo water-use study, primarily on the basis of the diversity of environmental and agricultural conditions under which wetland kalo is grown and landowner permission and availability. Flow and water-temperature data were collected at the lo`i complex level and at the individual lo`i level. To ensure that flow and temperature data collected at different lo`i reflect similar irrigation conditions (continuous flooding of the mature crop), only lo`i with crops near the harvesting stage were selected for water-temperature data collection. The water need for kalo cultivation varies depending on the crop stage. In this study, data were collected during the dry season (June-October), when water requirements for cooling kalo approach upper limits. Flow measurements generally were made during the warmest part of the day, and temperature measurements were made every 15 minutes at each site for about a two-month period. Flow and temperature data were collected from kalo cultivation areas on four islands - Kaua`i, O`ahu, Maui, and Hawai`i. The average inflow value for the 19 lo`i complexes measured in this study is 260,000 gallons per acre per day, and the median inflow value is 150,000 gallons per acre per day. The average inflow value for the 17 windward sites is 270,000 gallons per acre per day, and the median inflow value is 150,000 gallons per acre per day. The average inflow value for the two leeward sites is 150,000 gallons per acre per day. The average inflow value measured for six individual lo`i is 350,000 gallons per acre per day, and the median inflow value is 270,000 gallons per acre per day. The average inflow value for the five windward lo`i is 370,000 gallons per acre per day, and the median inflow value is 320,000 gallons per acre per day. The inflow value for the one leeward lo`i is 210,000 gallons per acre per day. These inflow values are consistent with previously reported values for inflow and are significantly higher than values generally estimated for water consumption during kalo cultivation. These measurements of inflow are important for future considerations of water-use requirements for successful kalo cultivation. Of the 17 lo`i complexes where water inflow temperature was measured, only 3 had inflow temperatures that rose above 27 ?C, the threshold temperature above which wetland kalo is more susceptible to fungi and associated rotting diseases. The coldest mean inflow temperature was 20.0 ?C and the warmest inflow temperature was 24.9 ?C. All 15 of the sites where outflow temperatures were measured had some temperatures greater than 27 ?C. Outflow temperatures exceeded 27 ?C between 2.5 percent and about 40 percent of the time. Mean outflow temperatures ranged from 23.0 ?C to 26.7 ?C.

Open-File Report↗

USGS2007

Natural and diverted low-flow duration discharges for streams affected by the Waiahole Ditch System, windward O`ahu, Hawai`i

For nearly a century, the Waiahole Ditch System has diverted an average of approximately 27 million gallons per day of water from the wet, northeastern part of windward O`ahu, Hawai`i, to the dry, central part of the island to meet irrigation needs. The system intercepts large amounts of dike-impounded ground water at high altitudes (above approximately 700 to 800 ft) that previously discharged to Waiahole (and its tributaries Waianu and Uwao), Waikane, and Kahana Streams through seeps and springs. Diversion of this ground water has significantly diminished low flows in these streams. Estimates of natural and diverted flows are needed by water managers for (1) setting permanent instream flow standards to protect, enhance, and reestablish beneficial instream uses of water in the diverted streams and (2) allocating the diverted water for instream and offstream uses. Data collected before construction of the Waiahole Ditch System reflect natural (undiverted) flow conditions. Natural low-flow duration discharges for percentiles ranging from 50 to 99 percent were estimated for four sites at altitudes of 75 to 320 feet in Waiahole Stream (and its tributaries Waianu and Uwao Streams), for six sites at altitudes of 10 to 220 feet in Waikane Stream, and for three sites at altitudes of 30 to 80 feet in Kahana Stream. Among the available low-flow estimates along each affected stream, the highest natural Q50 (median) flows on Waiahole (altitude 250 ft), Waianu (altitude 75 ft), Waikane (altitude 75 ft), and Kahana Streams (altitude 30 ft) are 13, 7.0, 5.5, and 22 million gallons per day, respectively. Q50 (median) is just one of five duration percentiles presented in this report to quantify low-flow discharges. All flow-duration estimates were adjusted to a common period of 1960-2004 (called the base period). Natural flow-duration estimates compared favorably with limited pre-ditch streamflow data available for Waiahole and Kahana Streams. Data collected since construction of the ditch system reflect diverted flow conditions, which can be further divided into pre-release and post-release periods - several flow releases to Waiahole, Waianu, and Waikane Streams were initiated between December 1994 and October 2002. Comparison of pre-release to natural flows indicate that the effects of the Waiahole Ditch System diversion are consistently greater at lower low-flow conditions (Q99 to Q90) than at higher low-flow conditions (Q75 to Q50). Results also indicate that the effects of the diversion become less significant as the streams gain additional ground water at lower altitudes. For Waiahole Stream, pre-release flows range from 25 to 28 percent of natural flows at an altitude of 250 feet and from 19 to 20 percent at an altitude of 320 feet. For Waikane Stream, pre-release flows range from 30 to 46 percent of natural flows at an altitude of 10 feet and from 7 to 19 percent at an altitude of 220 feet. For Kahana Stream, pre-release flows range from 65 to 72 percent of natural flows at an altitude of 30 feet and from 58 to 71 percent at an altitude of 80 feet. Estimates of post-release flows were compared with estimates of natural flows to assess how closely current streamflows are to natural conditions. For Waianu Stream, post-release flows at an altitude of 75 feet are 41 to 46 percent lower than corresponding natural flows. For Waikane Stream, post-release flows at an altitude of 75 feet are within 12 percent of the corresponding natural flows. Comparisons of pre-release and post-release flows for Waikane Stream at altitudes of 10 to 220 feet were used to assess downstream changes in flow along the stream reach where flow releases were made. For a particular stream altitude, proportions of pre-release to post-release flows associated with median flows are consistently greater than proportions associated with lower low flows because the relative effect of the flow release is smaller at higher low flows. Similarly, for a particular f

Hawaii↗

USGS2005

Rainfall-Runoff and Water-Balance Models for Management of the Fena Valley Reservoir, Guam

The U.S. Geological Survey's Precipitation-Runoff Modeling System (PRMS) and a generalized water-balance model were calibrated and verified for use in estimating future availability of water in the Fena Valley Reservoir in response to various combinations of water withdrawal rates and rainfall conditions. Application of PRMS provides a physically based method for estimating runoff from the Fena Valley Watershed during the annual dry season, which extends from January through May. Runoff estimates from the PRMS are used as input to the water-balance model to estimate change in water levels and storage in the reservoir. A previously published model was calibrated for the Maulap and Imong River watersheds using rainfall data collected outside of the watershed. That model was applied to the Almagosa River watershed by transferring calibrated parameters and coefficients because information on daily diversions at the Almagosa Springs upstream of the gaging station was not available at the time. Runoff from the ungaged land area was not modeled. For this study, the availability of Almagosa Springs diversion data allowed the calibration of PRMS for the Almagosa River watershed. Rainfall data collected at the Almagosa rain gage since 1992 also provided better estimates of rainfall distribution in the watershed. In addition, the discontinuation of pan-evaporation data collection in 1998 required a change in the evapotranspiration estimation method used in the PRMS model. These reasons prompted the update of the PRMS for the Fena Valley Watershed. Simulated runoff volume from the PRMS compared reasonably with measured values for gaging stations on Maulap, Almagosa, and Imong Rivers, tributaries to the Fena Valley Reservoir. On the basis of monthly runoff simulation for the dry seasons included in the entire simulation period (1992-2001), the total volume of runoff can be predicted within -3.66 percent at Maulap River, within 5.37 percent at Almagosa River, and within 10.74 percent at Imong River. Month-end reservoir volumes simulated by the reservoir water-balance model for both calibration and verification periods compared closely with measured reservoir volumes. Errors for the calibration periods ranged from 4.51 percent [208.7 acre-feet (acre-ft) or 68.0 million gallons (Mgal)] to -5.90 percent (-317.8 acre-ft or -103.6 Mgal). For the verification periods, errors ranged from 1.69 percent (103.5 acre-ft or 33.7 Mgal) to -4.60 percent (-178.7 acre-ft or -58.2 Mgal). Monthly simulation bias ranged from -0.19 percent for the calibration period to -0.98 percent for the verification period; relative error ranged from -0.37 to -1.12 percent, respectively. Relatively small bias indicated that the model did not consistently overestimate or underestimate reservoir volume.

Scientific Investigations Report↗