In Hawaii, ground-water recharge is commonly estimated with annual or monthly water budgets. Because monthly water budgets account for seasonal variability in rainfall and evapotranspiration, monthly water budgets generally provide more accurate recharge estimates than annual water budgets. Similarly, because daily water budgets account for daily variations in rainfall and evapotranspiration, daily water budgets generally provide more accurate recharge estimates than monthly water budgets. For this study, (1) ground-water recharge in the Hawi area was estimated using a daily water budget and compared to recharge previously estimated using a monthly water budget (Shade, 1995), and (2) ground-water availability was estimated and compared to a previous estimate of ground-water availability (Underwood and others, 1995).

The purpose of this report is to describe the (1) calculation of a daily water budget to estimate average annual ground-water recharge for 1990’s land-use conditions in the Hawi area of the island of Hawaii, (2) uncertainty in the recharge estimate, and (3) results from numerical ground-water flow models that simulate the hydrologic effects of additional ground-water withdrawals at rates between 10 and 20 million gallons per day (Mgal/d) above average 1990’s rates. No new data were collected as part of this study; only existing information was used to compute the water budget. An existing numerical ground-water flow model (Underwood and others, 1995) formed the basis for models used in this study to simulate the effects of additional with-drawals.

The Hawi study area is located on the windward (northeastern) side of the crest of the Kohala Mountains. The Kohala Mountains are formed by the Kohala Volcano, the oldest and northernmost of five volcanoes forming the island of Hawaii. The study area covers about 55 square miles and is bounded on the southwest by the crest of the Kohala Mountains, on the east by the eastern drainage divide of Pololu Stream, and on the north by the coast (fig. 1). Within the study area, the land-surface altitude ranges from sea level at the coast to about 4,000 ft near the headwater of Pololu Stream. In general, the land surface is moderately dissected. The dominant land cover is pasture, with smaller areas used for agriculture, commonly orchards, and rural and urban development (fig. 2). The upland area is covered in places with native forest vegetation (Jacobi, 1989). From the early 1900’s to the early 1970’s, sugarcane was grown over much of the area that is currently in pasture.

Mean annual rainfall in the Hawi area ranges from less than 40 in. near the coast at Upolu Point to between 120 and 160 in. inland, near the headwater of Pololu Stream (fig. 3). The rainfall distribution is controlled primarily by topography and wind direction. Persistent northeasterly winds, known locally as tradewinds, are forced up the slopes of the Kohala Mountains. The warm, moisture-laden air is orographically lifted and cooled, which frequently results in cloud formation and rainfall. Because the air commonly loses moisture as it flows over the Kohala Mountains, the area on the southwestern, leeward side of the mountain crest is drier, with less than 10 in. of annual rainfall in some coastal areas south of the Hawi study area.

Fog water that is intercepted by vegetation and subsequently drips to the ground is referred to as fog drip. Fog drip can exceed rainfall during some periods and contribute to recharge. For example, over a 100-day period at an altitude of 3,800 ft on the Kohala Mountains, Juvik and Nullet (1995) measured 23.9 in. of canopy throughfall (which includes fog drip), but only 13.2 in. of rainfall. Therefore, it can be inferred that about 10.7 in. of fog drip was collected during the 100-day period.

Davis and Yamanaga (1963) indicated that there are no perennial streams in the study area west of Pololu Stream, although data from continuous-record gaging stations are not available to characterize streamflow in the area. Presley (1999) measured flow at various sites on Pololu Stream following a period of dry weather in 1996 and indicated that Pololu Stream was dry along its entire length except in two places: where water from the Kohala ditch was leaking into an unnamed eastern tributary that flows into the main channel, and in a wetland area near the ocean.