1. Goals of ALOS Research Plan

To achieve the ALOS mission, it is essential not only to distribute data products, but also to promote scientific and utilization researches for ALOS data in broad categories ranging from the environmental and resource sciences to computer science. The ALOS research plan is made to achieve goals of calibration and validation of each sensor, scientific research, and utilization research promotion.
We define development of specific data products and algorithms for promoting the other scientific researches as "strategic goals." These are selected considering the relevance to the ALOS mission and the goals of this plan, resource limitations, etc. General goals rather than the strategic goals are mainly promoted through Research Announcements; the Earth Observation Research Center will be responsible for achieving the strategic goals.

2. Calibration and Validation of Each Sensor and Related Basic Studies

Calibration and Validation of PRISM, AVNIR-2 and PALSAR on processing Level 1 data from Level 0 data are most important and necessary to improve the accuracy of high resolution DEM and biomass distribution data. Moreover, related basic studies required for Calibration and Validation of these sensors are essential for development coming generation high performance sensors which have high performance.

3. General Goals

The general goals determine which categories to select, how to contribute to each category, and what kinds of data products and algorithms are required. The categories mentioned below are classified based on the categories of undergoing core projects of the International Geosphere- Biosphere Program (IGBP).

3.1 Land Use and Land Cover Research
This research reveals land use and land cover changes, and contributes to clarifying the mechanism of such changes and the development of change models. It is important to develop the following products and algorithms for these purposes.

3.1.1 High-resolution Digital Elevation Model
Topographical conditions strongly influence land use determination and its change process as well as environmental impacts such as soil erosion and runoff changes. In these research categories, a Digital Elevation Model (DEM) which corresponds to a 1: 25,000 to 1: 100,000 scale topographical map is useful. Algorithms for stereo matching and interferometric measurement need to be developed.

3.1.2 Orthophoto image (PRISM, AVNIR-2, PALSAR images) and land use and land cover data
These can reveal sprawl of urban areas and villages, changes of agricultural land and agricultural practices, deforestation, etc. Radar images may also be able to detect tillage variations (variation of tillage surface roughness) and changes of cropping pattern. It is also necessary to promote research for integrating ALOS data with ADEOS-II data.

3.2 Topography and Geology
This research contributes to measuring changes in terrain and watercourses due to soil erosion and slope failure as well as to classifying and analyzing terrain features with elevation data. It is thus essential that the following data products and algorithms be developed.

3.2.1 High-resolution DEM
High-resolution DEM can be used for terrain classification and analysis as well as watercourse analysis.

3.2.2 Orthophoto image (particularly PALSAR image)
An orthophoto image can be used for extraction and classification of terrain features and so on.

3.2.3 Elevation change due to soil erosion and sedimentation
Interferometric measurement is expected to provide a method for measuring time-series changes of land elevation. An area which a topographic condition changes remarkably due to soil erosion and sedimentation, such as the Yellow River basin, is selected as the objective area.

3.3 Terrestrial (Vegetation) Ecosystem , Agriculture and Forestry Research
This research contributes to clarifying vegetation dynamics with emphasis on the carbon cycle, monitoring agricultural production, estimating productivity of pastures based on the vegetation dynamics, and investigating biomass changes caused by human activities. For this purpose, the following data products and algorithms need to be developed using AVNIR-2 data or other satellite data.

3.3.1 Forest distribution monitoring
Methods for measuring global forestry distribution are expected to be advanced using PALSAR or AVNIR-2.

3.3.2 Vegetation biomass distribution measurement
Vegetation biomass is a key parameter which describes vegetation dynamics. A method of measuring vegetation biomass with focus on forests with simultaneous observations by PRISM and AVNIR-2 is expected to be developed.

3.3.3 Application to forest management
A method of monitoring deforestation and afforestation and estimating forest growth should also be developed concurrently with the development of a biomass measurement method.

3.3.4 Monitoring the productivity of pastures and crop land
Developing a method for determining the crop planting area, estimating productivity of pastures and crop land in a specific area, based on intensive observation by both PALSAR and AVNIR-2, is expected. In addition, a method of monitoring the changes of agricultural production and productivity of pastures caused by drought should also be developed.

3.3.5 Monitoring vegetation change due to human activities such as biomass burning
A method for measuring and monitoring the variation of biomass density and vegetation structure due to biomass burning in specific areas with intensive observations using PALSAR together with AVNIR-2 needs to be developed.

3.3.6 Desertification Monitoring
This aims at monitoring the decline of land productivity and soil degradation due to excessive cultivation and pasturage and improper irrigation. Methods of indirectly monitoring desertification need to be developed by observing vegetative deterioration using PALSAR and AVNIR-2 as well as directly monitoring of salt accumulation on the soil surface using AVNIR-2.

3.4 Climatic System, Hydrological Processes, and Water Resources Related Research

3.4.1 Surface process
In research on surface processes, it will be useful to develop methods to understand vegetation distribution, to measure soil moisture, and to prepare soil moisture datasets.

(1) Vegetation monitoring:
Development of algorithms for measuring key parameters for water vapor estimation such as biomass density or Leaf Area Index (LAI) is expected. Development of methods for integrating other satellite data, such as ADEOS-II data, is also important.
(2) Estimating of soil moisture distribution:
Development of algorithms for measuring soil moisture with PALSAR need to be facilitated. Development methods for integrating other satellite data, such as ADEOS-II data, with PALSAR data may also be essential.
(3) Run-off analysis:
ALOS data will contribute to run-off analysis under various conditions related to climate and land even in areas where there is insufficient available data.
l High-resolution DEM: A high-resolution DEM, having much higher resolution than the existing 1km DEM, has the potential of making the run-off analysis more accurate and reliable.
l Datasets of land use / land cover and their changes: These datasets will help analyze water valance and run-off variation due to land use and land cover changes. Using additional satellite data will make this research more successful.

3.4.2 Water pollution analysis
This research aims at estimating the quantity of water pollutant load and analyzing flow-down conditions by providing more accurate topographical data, and land use and land cover datasets.

(1) High-resolution DEM:
A high-resolution DEM will enable more accurate analysis of the flow-down of the water pollutant load due to soil erosion and estimation of the amount.
(2) Datasets of land use / land cover and their change:
These datasets facilitate analyzing the quantity of the water pollutant load by land use and land cover changes. Combined with hydrological analysis, these datasets reveal the condition of the pollution effluent. Using additional satellite data will make this research more successful.

3.4.3 Snow and ice related analysis
Accurately analyzing snow and ice in the following categories using high resolution sensor data from ALOS will contribute to understanding changes of climate and water resources (hydrological cycles), and so on.

(1) Estimating states and changes of snow cover and snow-water equivalent:
Analysis using the observation data from PALSAR and AVNIR-2 can help accurately predict and understand the seasonal or annual change of snow cover and snow-water equivalent.
(2) Measuring and analyzing variations of ice sheets and glaciers:
Analysis of Interferometric measurements by PALSAR and observation by AVNIR-2 will contribute to understanding the ice sheet mass balance and mountain glacier variation in the South Pole, Greenland, and so on.
(3) Sea ice monitoring:
Analyzing the observation data from PALSAR and AVNIR-2 will contribute to determining the extent and seasonal or annual variation of ice sheets in the polar regions and coastal zones. Furthermore, using ScanSAR data from PALSAR will contribute to methodological development of extensive sea ice monitoring, and using polarimetric data of PALSAR will improve the accuracy of sea ice classification.

3.5 Oceanography and Coastal Zone Related Research

3.5.1 Coastal zone related research
Providing information on wave, sea surface wind, water current, sea ice, topographical change and sand drift in coastal areas can support economic activities in coastal areas such as sea traffic, pollution control and fisheries. For this purpose, it is necessary to develop and prepare the following algorithms and products.

(1) Oil spill datasets of coastal zones:
Techniques for extracting the polluted areas from PALSAR images is expected to be developed. It is necessary to analyze sea surface wind and the spectrum of ocean waves around the area to accurately extract polluted areas. At the same time, datasets which analyze these factors must be developed.
(2) High-resolution DEM of coastal zones:
High-resolution DEM of coastal zones combined with water depth data will contribute to analyzing transformation of sea wave and coastal topography and impacts of sea level rise.
(3) Datasets of sea surface wind and wave height in coastal zones:
It is possible to prepare datasets for coastal sea-surface winds and waves using PALSAR data. A method which predicts coastal current by utilizing a numeric simulation model along with these datasets should be also developed. These are useful for giving of a boundary condition for analysis of coastal transformation and sand drift.
(4) Datasets of sea ice:
Methods for monitoring coastal sea ice and for providing its data accurately using PALSAR and AVNIR-2 need to be developed. Coastal ice datasets are useful for various coastal activities of human beings.

3.5.2 Ocean dynamics
Utilization of PALSAR or development methods using PALSAR together with other satellite data such as ADEOS-II data will contribute to studies on air-sea interaction, sea waves, and dynamics of various ocean phenomena in coastal zones and the open seas.

(1) Coastal topography-air-sea interaction:
Strong or weak wind zones are generated locally in a coastal sea because of coastal topography. Though such changes of sea-surface are essentially important to coastal waves and water currents, little research has been conducted in these areas. High-spatial resolution information collected by PALSAR on ocean waves and sea surface winds is expected to greatly contribute to studying the coastal topography-air-sea interaction and probing its mechanism.
(2) Wave-current interaction and various phenomena in the ocean:
Studies on the interactions between ocean waves and currents using data acquired in the ScanSAR mode of PALSAR need to be promoted. Based on these studies, large-scale ocean currents (like the Black Current), cold/warm water masses, coastal water currents, and internal waves can be visualized from ScanSAR images. This will help us to understand ocean dynamics.

3.6 Disasters and Earthquakes

3.6.1 Diastrophism
Methods for monitoring land surface deformations due to diastrophism employing interferometoric observation by PALSAR are needed to be developed.

3.6.2 Volcano monitoring
A method for monitoring deformation of mountains caused by volcanic activities should be developed.

3.6.3 Slope failure
It is necessary to develop a method for risk analysis of slope failure using high-resolution DEMs generated by PRISM and PALSAR. Datasets of land use and land cover in slope areas will contribute to estimating surface erosion and water infiltration as well as forecasting the damage of slope failure.

3.6.4 Analysis and simulation of flooding and inundation
By applying high-resolution DEMs, we can conduct run-off (flooding) analysis and inundation in areas where we previously haven't had enough data. This will contribute to advancing methods for analyzing and investigating those phenomena. At the same time, land cover and land use data will improve the reliability of these analyses as well as damage forecasting and refuge planning.

3.6.5 Tidal wave analysis
It is expected that tidal wave tracing analysis with high-resolution DEMs can be conducted in areas where we previously haven't had enough data. This will contribute to advancing the methods of analyzing and investigating these phenomena. Furthermore, land cover and land use data together with high-resolution DEMs will improve the reliability of these analyses as well as damage forecasting and refuge planning.

3.6.6 Disaster monitoring technique
Disaster monitoring techniques reveal damage due to drought, flood, fire, slope failure, earthquake disaster. Furthermore, these techniques can be applied to quick and accurate damage assessment (for example, the effect on agricultural production).

3.7 Resource Exploration
Resource exploration research techniques for mineral resource need to be developed. Analysis methods integrating PRISM, AVNIR-2, and PALSAR images with DEMs will be examined.

3.8 Development of Spatial Data Infrastructure

3.8.1 Techniques for developing spatial data infrastructure:
Automatic recognition and three-dimensional measurement of terrain features need to be developed to efficiently generate high-resolution DEMs and spatial data on artificial structures, which are the basis of various scientific research and practical uses. For three-dimensional measurement, orientation methods and stereo matching methods for PRISM images need to be developed. Furthermore, an algorithm for interferometric measurement need to be developed for PALSAR. In addition, a method integrating images (from PRISM, AVNIR-2 and PALSAR) with DEM needs to be developed for automatic recognition and three-dimensional measurement of terrain features such as roads, large structures and urban areas.

3.8.2 Management and retrieval techniques for very large database
Using ALOS data as a test case, techniques for very large spatial database are expected to be developed. Examples include data storage and management techniques, an efficient retrieval method based on a map or coordinates.

3.9 Basic Studies on Scattering and Interferometric Characteristics
In order to expand the application fields of PALSAR data, including improvements of interferometric analysis, polarimetric analysis, and terrain correction methods, the following study will be performed.

3.9.1 Decomposition method for polarimetric SAR data
Decomposition methods for PALSAR polarimetric data should be studied and developed. This methodology will be applied to land cover classification using scattering characteristics of the targets.

3.9.2 Polarimetric and interferometric data analysis
Interferometric analysis is applied to the polarimetric data acquired from PALSAR repeat-pass observation. An applied field example is tree height estimation in forested areas.

3.10 Basic studies for accurate observation with high resolution optical sensors
Research on the following topics needs to be conducted to develop the next-generation high-resolution optical sensors.

(1)The accuracy of satellite position and attitude determination, including the rate of the variation of the attitude which will affect the pointing accuracy and resolution of the optical sensors, needs to be analyzed and evaluated.
(2)Impacts of the shock during launch, temporal degradation, and temperature changes inside the instruments on optical alignment (including the optical benches and the structures with optical alignment), photoelectric transfer characteristics, and sensor resolution need to be analyzed and evaluated.
(3)It is necessary to develop a code to analyze the effect of multiscattering of the atmosphere, especially regarding aerosols, whose spatial conditions fluctuate largely with time, and to estimate the surface albedo with high speed and high accuracy.
(4)A suitable filter for the modulation transfer function (MTF) correction needs to be developed to restore observation data degraded by the MTF of each sensor or atmospheric influences.

4. Strategic Goals

We define development of specific data products and algorithms for promoting the other scientific researches as "strategic goals." These are selected considering the relevance to the ALOS mission and the goals of this plan, resource limitations, etc.

4.1 Data products

4.1.1 Global High-resolution DEM and Orthophoto image (PRISM, AVNIR-2, and PALSAR)
These data products form the basis of many fields of research and practical applications. They are provided by only ALOS at the moment. However, resources required to generate these data are so large that the accuracy and resolution may change according to the objective area. Global coverage will be pursued by coordinating with other data node organizations.

4.1.2 Global Biomass density dataset (PALSAR and AVNIR-2)
Biomass is not only one of the most important parameters for estimating the carbon cycle, but also provides a basis for forestry management. However, it is difficult to measure on the ground and there is no data covering a large area. Since only ALOS is equipped with L-band, which favors biomass observation, it is expected that biomass density data will be generated using PALSAR images along with AVNIR-2 images and high resolution DEMs. These activities will allow us to conduct time series analysis with Global Forest Mapping (GFM) datasets from JERS-1 SAR data.

4.1.3 Land surface deformation dataset (Earthquake-prone areas only)
The distribution of deformed land surfaces can be extracted by interferometric measurement. Monitoring diastrophism is essential in the Pacific Rim area, including Japan, which is always threatened by earthquakes. Land surface deformation data will be collected by periodic satellite observation and continuous ground observation.

4.2 Algorithms

4.2.1 Automated generation of high-resolution DEM and orthophoto image
A large computing capability is usually required to generate high-resolution DEMs and orthophoto images, and the quality of these products is affected strongly by the performance of the algorithms used. Algorithms for automated generation of high-resolution DEMs and orthophoto (including an algorithm to estimate satellite position and altitude) need to be developed.

4.2.2 Accuracy improvement of biomass measurement method
Development of algorithms using DEMs and AVNIR-2 images together with other satellite images for measuring global biomass distribution with higher accuracy is solicited.

4.3 Calibration and Validation for each Sensor and Related Basic Studies
Calibration and validation of each sensor is necessary for improviing the quality of the data products such as high-resolution DEMs and biomass density data. In addition, basic studies on calibration and validation for improving the accuracy of each sensor should also be pursued as strategic goals.

4.3.1 Calibration and validation for optical Sensors
To generate high-quality products from optical sensors, AVNIR-2 and PRISM, basic study for very accurately evaluating radiance characteristics, geometric characteristics, spatial resolution, system noises, and other factors. is considered to be one of the strategic objectives.

(1) Accuracy improvement of radiance and brightness calibration
The radiance and brightness of optical sensors will be calibrated by using pre-flight test data, internal calibration source data, and external calibration data after launch. The main output of this study is to estimate absolute calibration coefficients. In particular, an important challange will be the improvement of stability characteristics with ground-based experiments with calibration after launch and development of the radiative transfer model with high accuracy.
(2) Accuracy improvement of DEM
Algorithms for automatically evaluating and correcting registration and pointing accuracy, and for automatically producing high-resolution DEMs using stereo matching images will be developed.
(3) Atmospheric correction
Algorithm should be improved to estimate the surface albedo on a heterogeneous surface using optical sensors data, taking into account the effect of multi-scattering in the atmosphere, especially spatial and temporal changes of aerosols.

4.3.2 Calibration and validation for PALSAR system
A basic study for achieving high radiometric accuracy of the PALSAR system is considered to be one of the strategic objectives.

(1) Accurate estimation of normalized radar cross section
The relation between the digital number and the normalized backscattering coefficient for PALSAR standard products will be determined by using the pre-flight test data, internal calibration source data, and external calibration data. The main outputs of this study are the estimated in-orbit antenna elevation patterns and the absolute calibration coefficients.
(2) Accuracy improvement of interferometric SAR data
In order to derive accurate digital elevation models as well as crustal movements, a study on achieving an accurate phase difference will be done by using repeat-pass interferometric datasets acquired by the PALSAR system.
(3) Accuracy improvement of polarimetric SAR data
PALSAR's polarimetric observation mode is currently an experimental mode. However, this observation mode will be the main operation mode in future SAR systems. In order to prepare for the practical use of fully polarimetric data, polarimetric calibration with the data acquired from PALSAR polarimetric observation mode should be studied. The methodology to derive phase correction, cross talk, and gain imbalance will be developed and investigated.