Tropical dry forests are among the most endangered and degraded ecosystems in the world. There is a great concern that Hawaii's remaining dry forest ecosystems could vanish without intensive, aggressive management and restoration efforts due to three interrelated factors: grazing, fire, and alien grass invasion. The purpose of this project is to examine the effectiveness of remote sensing and GIS in delineating areas which require intensive management.

FUNDING has been provided to CTAHR for this research from the USDA Cooperative State Research, Education, and Extension Service Project HAW00124-M.

OBJECTIVES

We propose to conduct research to improve our ecological understanding of the dry forest decline process and restoration potential through an integrated use of geospatial data (remote sensing and geographic information systems, GIS) and detailed field measurements at a region-wide scale. We propose to use the ahupuaa of Puu Waawaa in the North Kona region as our study area, one of the largest areas of remaining dry forest habitat in the state. The proposed work has the following objectives:

1) To characterize the historical spatial patterns of dry forest cover and their changes across the entire ahupuaa of Puu Waawaa using time-series airphoto data sets covering a period of 50 years, from 1954 to 2001;

2) To characterize the current state of forest health by combining field-measured parameters of forest health conditions with advanced, satellite remote sensing imagery (1m to 30 m spatial resolutions) and;

3) To establish linkages between forest health and decline to both the natural and human-induced environmental conditions of soils, substrate geology, climatic conditions, grazing, fire occurrences/ burned areas, and alien species invasions as well as historical land use information from recently-compiled, archival-historical documents (Maly, 1999; 2001).

APPROACH

We propose to conduct research to improve our ecological understanding of the dry forest decline process and restoration potential for the ahupuaa of Puu Waawaa in the North Kona region. Our approach will be: 1) to characterize the historical and current spatial patterns of dry forest cover and their changes using time-series airphotos, 2) to characterize the current state of forest health by an integrated use of advanced remote sensing data and field-based physiological measurements, and 3) to establish links of the historical land use practices (e.g., cattle ranching) and the environmental conditions with the current forest health. We will use time-series airphoto data sets covering a period of 50 years (1954 to 2001), recent high-spatial resolution remotely sensed images (1m to 30m), and a variety of geospatial, or geographic information system (GIS) data depicting the environmental conditions and land uses. Field-measurement data will include biophysical and biochemical indicators of forest health and vigor, species compositions, and the impact assessment results of pests and pathogens on forest restoration efforts. Comparisons will be made among intact dry forests, degraded forest remnants, and restored forest areas both temporally and spatially in order to understand the process of decline and the potential for restoration. All the geospatial data together with field-observational data will be integrated into a GIS database. This will result in a series of maps depicting the spatial variability of such forest conditions as forest health, degradation conditions, and spatial extent of invasive species within the Puu Waawaa ahupuaa. These will be used to identify and prioritize management and restoration actions, and to estimate a potential cost for various management/restoration scenarios. The goal of this research is to establish an assessment and evaluation system that can be used to develop cost-effective and scientific-sound, sustainable management plans for dry forest ecosystems throughout Hawaii and the tropics.

During this first year of the project period, we have had two accomplishments. 1) Characterization of current forest health using remote sensing and field data (Task 3 of the proposal). We obtained a set of IKONOS fine spatial resolution images over our study site, Puu Waawaa located in the north Kona region of the big island. The images were successfully processed using a new object-oriented classification algorithm and the dry forest tree cover information were successfully extracted with a good accuracy. We demonstrated the utility of the IKONOS sensor and the object-oriented classification algorithm in assessing ecosystem conditions. The resultant image (map) will be useful among foresters, resource managers, and federal/state agencies to monitor the effectiveness of various restoration methods being tested in Puu Waawaa. 2) In situ characterization of forest health and restoration potential (Task 5 of the proposal). Two field campaigns were conducted n the May-July period to assess the dry forest health conditions at selected locations in details (seven one-hectare plots). Percent cover of overstory trees, their species compositions, and structures (DBH and tree height) were measured and populated into a database. The data were successfully used to validate satellite estimates of tree cover.

We expect that the protocol newly-developed and tested in this project, that is, the use of fine-resolution satellite data classified using the object-oriented classification algorithm, will be a new standard methodology at federal/state agencies for producing various land use/cover maps and vegetation maps on an operational basis. We also expect that the protocol and maps generated will be used to monitor and assess the effectiveness of various management practices for restoring native dry forests and/or for the enhancement of State's natural resource base in a large-scale.