Satellite Technology for Mitigating Deforestation

Transcription

1 National Directorate of Forestry Ministry of Lands, Environment and Rural Development Mozambique Satellite Technology for Mitigating Deforestation Poland, Katowice, COP24 December 6 th 2018 Joaquim A. Macuácua GIS and RS Expert JAXA, Side Event Head of division of mapping and Data management

2 Content 1. General Information 2. Map elaboration fo land use and forest cover 3. Development of maps of reference of forest cover 4. Development of thematic maps 2 5. Identification of the deforestation areas by analysis of radar satellite images and evaluation of How to leverage information 6. ; 7. Identification of deforestation areas by analises radar images for the JJ-Fast 8. Conclusions

3 3 General information Area ~ Sq km; Population ~ (2017 census projection); Economy Agro-based (cashew nuts, cotton); Resources: Water, Wood Products, Shrimps, Natural Gas, Coal, Hydro-energy; Tropical climate with two seasons: wet season from October to March, and dry season from April to September; Institution responsible for Forest: National Directorate of Forestry under Ministry Land, Environment and Rural Development. Deforestation rate: 0.58%/year (NFI, Marzole 2007) with 41 Million ha of forest; Supported by JICA with 5 years project; Supported by JAXA in K&C#3 K&C#4 initiatives DINAF, 2018 based forest map

4 4 Objectives Design methodology for national forest monitoring system for REDD+; Ground Truth survey in unsurvey area of Plantations and natural forest; Obtain the samples data that are insufficient for the threshold setting; Compare the difference value -3, -4, -5 db between the images before and after 1 year Record the Non-forest area(thicket) in order to identify the threshold between forest/non-forest Ground Truth survey to support JJ-FAST system (analysed by ScanSAR imagery)

5 5 Activities of RS of 5 years JICA project Basic traing on IR Understanding of IR characteristics - Detection test of 2013 deforestation - Theory study - Radar images acquisition Training on job in Moz. and Jp. on IR - Ground survey for validation of the deforestation and linear value Detection of deforestation areas at national level - Evaluation how to make the use of the methodology - Visual and Automatic detection of deforestation Aplaied Practical training on IR Evaluation of the linear value - Analysis of precision - Water body and water strim Masking Traing in Japan in aplaied areas analysis

6 6 4. Identification of deforestation areas using radar image analysis and evaluation of wayforwad with the information Principles of dectation of Radar images Deforestation sites are detected by comparing radar images (ALOS-2) at two time series. The deforestation sites are those that indicate decrease of back-scattering (return) of radar waves. Therefor, higher return (forest) sites with light color and lower return (non-forest) dark color. Forest No Forest

7 7 4. Identification of deforested areas by radar image analysis and evaluation of the ways of using this information Determination of linear value (threshold) Sampling in Field (Ground Truth) The threshold values of temporal differences of the deforested areas were collected in 76 samples from Cabo Delgado, Niassa, Zambesi, Manica, Inhambane and Gaza. Study of the area with images support Deforested area Interview with local people

8 8 Yellow: GPS track White: GPS point North NS12 East 08/July/ m resolution 06/July/ m resolution West South - Deforestation for Maize. Past forest type is Mixed forest. - Deforestation area is 9.4 ha. - Big trees are deciduous (h=15m),lower trees are evergreen (Massuco, h= around 7m).

9 9 Recent Ground truth for JJ-Fast Calibration data in dray forest with Japan TDU and DINAF

10 10 4. Identification of deforested areas by radar image analysis and evaluation of the ways of using this information Identification of deforested areas by Radar Image Analysis Determination of the linear value to detect the deforested area db Big difference Small difference Dense forest db Forest open -16 db db -20 db Only deforestation data greater than 1.0 ha Beginning of the period End of period 10

11 4. Identification of deforested areas by radar image analysis and evaluation of the ways of using this information Result of areas of forest loss( ) 11 Province No. of sites indicating forest loss Deforested area (ha/year) Cabo Delgado Niassa Nampula Zambézia Tete Manica Sofala Inhambane Gaza Maputo Total

12 12 4. Identification of deforested areas by radar image analysis and evaluation of the ways of using this information How to leverage information Size of deforested area by forest type (CD): We identified the area deforested by forest type, from C. Delgado, overlapping the image with the 2008 forest coverage map. Deforeste d area (ha) (Semi-) dense ever green Semiopen ever green Mangrove (Semi-) deciduou s dense (Semi-) deciduous Open Total (ha) Deforestation in : 35,453 ha 1 ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ Total % of forest cover 0,81% 0,26% 0,01% 0,49% 0,34% 0,39% Map of forest cover and land use years during

13 13 4. Identification of deforested areas by radar image analysis and evaluation of the ways of using this information How to leverage information Size of deforested area by forest type (GZ): We identified the deforested area by forest type, from Gaza, overlapping the image with the forest cover map of Deforeste d area (ha) (Semi-) dense green (Semi-) dense ever dreen Mecruss e (Semi- ) dense decidu ous (Semi-) opnen decidu ous Mopan e Total (ha) 1 ~ Deforestation in : 5,342 ha 2 ~ ~ ~ ~ ~ ~ ~ ~ ~ Total % of forest cover 0,58% 0,21% 0,01% 0,09% 0,05% 0,17% 0,09% Map of forest cover and land use years during

14 Conclusions 1. The use map and forest cover was elaborated as a base map. 2. The reference forest cover maps were developed based on the maps above. 3. Thematic maps were also developed. 4. Deforested areas were identified by radar image analysis and ScanSAR the ways of using this information were evaluated It is difficult to survey the deforestation boundaries, then we do check the situation of deforestation areas and take overall pictures 6. Scanning deforested areas using drone is much efficient;

15 15