ISSN: 0852-0682.
E-ISSN: 2460-3945
Forum Geografi.
Vol 30 .
July 2016: 69-76 Radar Data for.
(Wuryant.
Radar Data for Identifying the Characteristics of Tropical Forest Stands Agus Wuryanta Balai Penelitian Teknologi Kehutanan Pengelolaan Daerah Aliran Sungai Corresponding author .
mail: agus_july1065@yahoo.
Abstract Radar is one of remote sensing technology which utilizes active electromagnetic energy and are able to provide information about the characteristics of forest stand.
This study utilized JERS-1 (Japanese Earth Resources Satellite-.
and ERS-1 (European Remote Sensing Satellite-.
radar images to analyze the relationship between the radar backscatter value and forest stand characteristics such as Diameter Breast Height (DBH), basal area, and canopy cover.
This research was conducted in Jambi Province.
Bungo Tebo District.
Sumatra.
Indonesia.
The research site covered the forest concession.
Suku Anak Dalam village, the area adjacent to Pelepat and Batang Tebo River, and Kuamang Kuning village.
Gamma Map Filter with 7 x 7 window size was applied to reduce speckle noise of the SAR images (ERS-1 and JERS-.
This study found out the positive significant correlation between basal area and DBH with JERS-1 radar backscatter value .
, r = 0.
75 and r = 0.
, while ERS-1 radar backscatter value has correlation .
= 0.
with the canopy cover.
Keyword: Radar, forest stand characteristics, backscatter value.
Abstrak Radar merupakan salah satu teknologi penginderaan jauh yang menggunakan energi elektromagnetik aktif dan memiliki kemampuan untuk mendapatkan informasi tentang karakteristik tegakan hutan.
Penelitian ini menggunakan citra satelit radar JERS-1 (Japanese Earth Resources Satellite-.
dan ERS-1 (European Remote Sensing Satellite-.
untuk menganalisis hubungan timbal balik antara nilai hamburan .
dan karakteristik tegakan hutan seperti diameter setinggi dada, basal area dan tutupan tajuk.
Penelitian dilaksanakan di Provinsi Jambi.
Kabupaten Bungo Tebo.
Pulau Sumatra.
Indonesia.
Lokasi penelitian meliputi wilayah Hak Pengusahaan Hutan (HPH), kampung Suku Anak Dalam dan berdekatan dengan sungai Pelepat dan Batang Tebo, serta Desa Kuamang Kuning.
Filter Gamma Map dengan ukuran 7 x 7 diterapkan pada citra radar (ERS-1 dan JERS-.
untuk mengurangi speckle noise.
Hasil penelitian menunjukkan adanya korelasi positif yang signifikan antara basal area yaitu r = 0,75 dan diameter setinggi dada yaitu r = 0,70 dengan nilai hamburan .
pada citra JERS-1, sedangkan nilai hamburan .
pada citra ERS-1 memiliki korelasi sebesar r = 0,64 dengan tutupan tajuk.
Kata Kunci: Radar.
Karakteristik Tegakan Hutan, nilai hamburan.
Introduction Indonesia is one among the countries having large tropical forests after Brazil and Zaire.
The major part of the Indonesian land surface, , 143 million ha .
%) has the status of forest land.
However, forests and peatlands in Indonesia is suffering from destructions (Riyanto.
Tropical forests are rich in trees and liana, in addition to herbaceous taxa.
Wet tropical forest, typically, has more species of herbs .
ncluding epiphyte.
or shrubs.
For the management and surveillance of forest product abundance, information on forest stand characteristics such as Diameter Breast Height (DBH), basal areal, and canopy cover are required.
In fact, the sustainable forest management is an important issue in Indonesia.
Forest management cannot be implemented without an understanding on the basic ecology of the forest.
One prerequisite for sustainable forest management is reliable information on the dynamics and characteristics of the stands since it is essential to know how the forest will grow and respond to natural condition or occasional disturbances (Riyanto and Wuryanta.
Extraction of information on the characteristics of forest stands from remotely sensory data is a fundamental activity Available online at http://Journals.
id/index.
php/fg/article/view/1501 Forum Geografi.
Vol 30 .
July 2016: 69-76 Radar Data for.
(Wuryant.
due to the necessity for a variety of applications including sustainable forest management According to international agreement (ITTO.
, the obligation of sustainable forest management for timber production must be achieved by the year of 2000.
Since 1989.
Indonesian Government through the Ministry of Forestry has been doing an inventory of forest by using aerial photo, landsat, and SPOT satellite data (Ministry of Forestry Republic of Indonesia.
However, the current methods were evidenced to be inaccurate in acquiring data on annual basis due to the frequent obstacle of severe cloud, fog, and rain.
Radar image can be used to overcome the problem.
Radar operates in a frequency band of the electromagnetic spectrum known as microwave region.
Most radar have been designed for a specific frequency or wavelength in which it is essential to understand the respective attributes of the different wavelengths in order to interpret an image (Trevett.
, 1.
According to Hord .
, the advantage of radar remote sensing is due to the wavelength region of radar .
5 cm and 1 .
that results on the capacity to penetrate clouds and smoke, even the wave longer than 2 cm can penetrate fog and precipitation.
Based on Sanden .
, the capacity of the radar to penetrate the forest canopy, especially the longer wavelength, also include the capability for distinguishing the primary and secondary forest type.
While at X-band .
75 cm to 5.
21 c.
and C-band .
cm to 7.
69 c.
, the zonation of vegetation, the identification of deciduous/coniferous forest, and the determination of age/height differences are rated as possible.
L-band .
cm Ae 76.
9 c.
appeared to be useful for the differentiation of some broad vegetation classes such as coniferous/deciduous, forest/ non-forest, and flooded/non flooded forest class categorization (Hoekman.
The main objective of this research was to analyze the relationship between the data of ERS-1 and JERS-1 radar backscatter value and the characteristics of forest stands i.
DBH, basal area, and canopy cover.
Research methods Location This research was conducted in coordinate between 01o15Ao00Ay to 01o45Ao00Ay South Latitudes and 102o15Ao00Ay to 102o45Ao00Ay East Longitudes.
This area lies in Jambi Province.
Bungo Tebo District.
Sumatra.
Indonesia.
addition, the site also covers the area of forest concession.
Suku Anak Dalam.
Pelepat and Batang Tebo River, and Kuamang Kuning Village.
Research location is presented in Fig.
Figure 1.
Research location Forum Geografi.
Vol 30 .
July 2016: 69-76 Radar Data for.
(Wuryant.
Materials and Tools The materials used in this research were radar satellite images in digital format, topographic map, and thematic maps.
The radar satellite images included the European Remote Sensing satellite-1 (ERS-.
and the Japanese Earth Resources Satellite-1 (JERS-.
ERS-1
satellite operates several sensors, including SAR (Synthetic Aperture Rada.
and ATSR (Along-Track Scanning Radiomete.
, while JERS-1 satellite is land resource satellite which operates radar sensor together with optical sensor (Danoedoro P.
This research utilized multi-temporal ERS-1 radar data obtained on January 14, 1995 (ERS-.
May 04, 1995 (ERS-.
, and June 08, 1995 (ERS-.
The first number indicated the generation of the satellite (ERS-.
while the second was intended to differentiate the scenes from one to each other.
The characteristics of ERS-1 and JERS-1 are presented in Table 1.
Table 1.
ERS-1 and JERS-1 image characteristics.
Parameters Frequency Wavelength
Type
Band width
Antenna
Polarization Angle of Incidence Height of orbit
Track width
Range resolution Azimuth resolution Recurrent day
ERS-1
C-band .
3 GH.
7 cm Chirp radar.
SAR 5 MHz Length 10 m, width 1 m 785 km 80 Ae 100 km 35 days JERS-1 L-band .
2 GH.
5 cm
SAR
15 MHz Length 12 m, width 2.
568 km 75 km 18 m .
44 days Source: https://directory.
org/web/eoportal/satellite-missions/e/ers-1 (European Remote Sensing Satellite ERS-.
https://directory.
org/web/ eoportal/satellite-missions/j/jers-1.
There were four topographic maps used in the research area, namely.
Lubuk Punggai.
Muara Ketato.
Muara Kilis, and Muara Tebo.
The scale of topographic maps was 1:50,000.
While the thematic map used in this research was soil map with a scale of 1:200,000.
The tools included digital image processing and GIS analysis such as hardware and software (Erdas Imagine version 9.
1 and ArcGIS 9.
Field survey equipment consisted of compass.
Global Positioning System (GPS), ballpoint, and clipboards.
Printing equipment consisted of cartridges and papers.
Methods According to Lillesand and Keifer .
, raw digital images usually endure significant geometric distortions that thus they cannot be used as maps.
In addition, radar image has unique characteristics in compared with optical image because of its speckle noise.
Gamma Map Filter with 7 x 7 window size was applied to reduce speckle noise of the SAR images (ERS-1 and JERS-.
Geometric correction converted a raster image after filtering as slave to a topographic map reference in a metric as Clearly identifiable points on both the slave and master map were selected for Affine transformation was applied to transform the image.
Selection of sample plots was done using simple random sampling Coordinates of the plot .
nown as centre of the plo.
were then obtained from Landsat Thematic Mapper (TM) image and topographic map.
Sample plots were collected during field work activities.
To ensure each centre of the plot, the coordinates were measured using Global Positioning System Forum Geografi.
Vol 30 .
July 2016: 69-76 Radar Data for.
(Wuryant.
(GPS).
After the centre of the plot was determined, a circular plot was measured out with a radius that would lead to a total area of 500 m2.
The radius would be in the range of approximately 12.
62 m to 12.
70 m.
Total number of sample plots were 58 sample plots:
45 sample plots on forest area and 13 sample plots on rubber plantation.
plots in which 58 sample plots were used in this research.
The parameters included the coordinates of the plot centre, slope, aspect, canopy cover, undergrowth cover and landuse.
In case of the plot was on a forest or rubber plantation, measurement of Diameter at Breast Height (DBH) greater than 10 cm was done in each tree.
From the dbh measurements, the Results and Discussion number of tree in the plot was recorded.
The average tree height was also estimated.
The Measurement and recording of plot result of measurement and recording of plot Plot parameters were measured and recorded parameters is presented in table 2.
on a prepared tally sheet.
There were 86 sample Table 2.
Sample plots coordinate and recording of plot parameters Coordinate Backscatter value Forest stand characteristics Trees Landuse JERS- ERS1
ERS13
ERSC.
DBH
BA/ha Heigh Forest Forest Forest Forest Forest Forest Forest Forest Forest Forest Forest Forest Forest Forest Forest Forest Forest Forest Forest Forest Forest Forest Forum Geografi.
Vol 30 .
July 2016: 69-76 Radar Data for.
(Wuryant.
Forest Forest Rubber Rubber Rubber Rubber Rubber Rubber Rubber Rubber Rubber Rubber Rubber Rubber Rubber Forest Forest Forest Forest Forest Forest Forest Forest Forest Forest Forest Forest Forest Forest Forest Forest The Relationship Between Radar Backscatter Forest Stand Characteristics in association with the Function of Wavelength.
Incidence Angle, and Polarization The influence of the wavelength on the radar backscatter is related to the capability of the radar energy in penetrating the forest canopy.
Longer wavelengths, such as L Aeband .
, have higher ability to penetrate the forest canopy than C-band .
7 c.
C-band has low penetration capability and the signal is reflected mainly by the canopy of the JERS-1 image showed a good positive correlation .
, r = 0.
with basal area and average DBH .
, r = 0.
Height of trees had insignificant correlation .
, r = 0.
with JERS-1 radar backscatter.
Canopy cover also showed insignificant correlation .
, r = 0.
with JERS-1 radar backscatter, indicating that longer wavelength such as L-band had significant relationship with forest stand characteristics such as DBH and basal area.
ERS-1 with C-band radar backscatter showed insignificant correlation with forest stand characteristics, except with canopy cover.
There was a considerable positive relationship between radar backscatter of ERS-11 image .
, r = 0.
with canopy cover.
While there was insignificant correlation .
, r = 0.
Forum Geografi.
Vol 30 .
July 2016: 69-76 Radar Data for.
(Wuryant.
between ERS-13 and canopy cover, moreover Estimation of Forest Stand Characteristics ERS-15 had positive correlation .
= 0.
Using Radar Backscatter Data of forest stand characteristics in this with canopy cover.
study was collected from the natural forest Polarization describes the orientation of the with the method of sampling technique.
electric field component of an electromagnetic Regression analysis was performed on the There are two types of polarization.
DBH, canopy cover, basal area, and average namely, like polarization and cross polarization.
height with the radar backscatter of ERS-1 JERS -1 was using VV polarization while and JERS-1 images.
Regression equations of ERS-1 was using HH polarization.
HH the relationship between radar backscatter and polarization means that the antenna of radar the forest stand characteristics are presented sends and received the signal in horizontal on Table 3.
Based on Table 3.
JERS-1 radar mode, while VV polarization means that the backscatter has positive correlation with antenna of radar sends and received the signal basal area and DBH, it means that basal area in vertical mode.
Different polarizations will and DBH can be estimated by using JERS-1 influence the obtained information of the radar image.
Correlation between basal area For instance, polarized data provides and backscatter value of JERS-1 radar image more information on the images especially is presented in Fig.
While Fig.
information on forest structure under the the correlation between average DBH and forest canopy as the result of the ground-trunk backscatter value of JERS-1 radar image.
interactions (Hoekman,1.
Figure 2.
The correlation between Basal Area per Hectare and JERS-1 Radar Backscatter value Forum Geografi.
Vol 30 .
July 2016: 69-76 Radar Data for.
(Wuryant.
Figure 3.
The correlation between Average DBH and JERS-1.
Radar Backscatter value Table 3.
Correlation model between radar backscatter of ERS-1 and JERS-1 with forest stand
Parameters Canopy cover
DBH
Average Height Basal Area Conclusions JERS-1 Y = 0.
0123x2 Ae 15x 150.
Y = -0.
Y = -0.
Y = -0.
019 x2 91x 96.
ERS-11
ERS-13
ERS-15
Y = 0.
017 x2 Ae Y = 0.
004x2 Ae 0.
25x Y = 0.
01x2 Ae 064x 29x 133.
Y = 0.
Y = -0.
Y = 0.
Y = -0.
Y = 0.
037x2 Ae 0.
Y = -0.
Y = -0.
Y = -0.
Y = 0.
Forest stand characteristics such as basal area and average DBH can be estimated JERS-1 radar backscatter values had by using JERS-1 radar image.
significant correlation .
, r = 0.
with basal area per hectare and average DBH .
= 0.
Height of trees and canopy cover Recommendations had insignificant correlation with JERS-1 1.
Despite the excellence of remote sensory images, it is advisable to have a prior radar backscatter values.
knowledge of the study area such as crop ERS-1 radar backscatter values had calendar and other indigenous practices insignificant correlation with forest stand that may affect the land use to obtain the characteristics such as average DBH, optimal result.
basal area, and average height This study found the best regression 2.
Remote sensing radar can be used in all weather conditions, particularly for forest equations that can be applied to estimate monitoring project, but it still has some basal area and average DBH was obstacles including .
Radar imagery developed from radar backscatter values system has totally different physically of JERS-1 image.
Forum Geografi.
Vol 30 .
July 2016: 69-76 Radar Data for.
(Wuryant.
and geometrically principles in compared with the optical systems .
The general lack of understanding on microwave interaction with vegetated terrain.
For some regions that are covered by clouds, rain and fog permanently, radar satellite imagery can be used as a substitute of optical images for forest monitoring and inventory Acknowledgement I would like to express my sincere gratitude and thank to Mr.
Reinink staff of IPL.
ITC The Netherlands who guide me during the image processing activity.
Also for Sam.
Hamid, and Belinda for the collaboration in collecting the data and discussion during the data processing.
References