Close

1. Identity statement
Reference TypeConference Paper (Conference Proceedings)
Sitemtc-m21d.sid.inpe.br
Holder Codeisadg {BR SPINPE} ibi 8JMKD3MGPCW/3DT298S
Identifier8JMKD3MGP3W34T/4AC8LC8
Repositorysid.inpe.br/mtc-m21d/2023/12.11.16.49
Metadata Repositorysid.inpe.br/mtc-m21d/2023/12.11.16.49.43
Metadata Last Update2024:01.02.17.16.56 (UTC) administrator
Secondary KeyINPE--PRE/
Citation KeyGanemXueShim:2023:MuReSe
TitleUnveiling Land Cover Changes in Northeast Brazil: A Multi-Source Remote Sensing Mapping Approach
Year2023
Access Date2024, May 24
Secondary TypePRE CI
2. Context
Author1 Ganem, Khalil
2 Xue, Yongkang
3 Shimabukuro, Yosio Edemir
Resume Identifier1
2
3 8JMKD3MGP5W/3C9JJCQ
Group1
2
3 DIOTG-CGCT-INPE-MCTI-GOV-BR
Affiliation1 University of California, Los Angeles (UCLA)
2 University of California, Los Angeles (UCLA)
3 Instituto Nacional de Pesquisas Espaciais (INPE)
Author e-Mail Address1
2
3 yosio.shimabukuro@inpe.br
Conference NameAGU FAll Meeting
Conference LocationSan Francisco, CA
Date11-15 Dec. 2023
PublisherAGU
Book TitleProceedings
History (UTC)2023-12-11 16:49:43 :: simone -> administrator ::
2024-01-02 17:16:56 :: administrator -> simone :: 2023
3. Content and structure
Is the master or a copy?is the master
Content Stagecompleted
Transferable1
Content TypeExternal Contribution
Version Typepublisher
AbstractNortheast Brazil (NEB) is a region heavily impacted by climate variability and land degradation, posing unique challenges for mapping and quantifying land cover changes (LCC). The high cloud incidence, strong phenological responses, and the influence of soil background reflectance on vegetation's spectral response have made accurate LCC assessment difficult in NEB. Additionally, existing products often overlook non-forest formations, leading to confusion in calculations. To address these issues, we conducted the first comprehensive pan-NEB LCC mapping from 2000 to 2020 using a multi-sensor-based methodology. Leveraging Google Earth Engine, we constructed high-quality, multitemporal cloud-free MODIS imagery composites, enhancing vegetation response, particularly in drought-prone ecosystems. To improve accuracy, Landsat and Sentinel-2 data were utilized to extract endmembers and apply the linear spectral unmixing (LSU) model. This estimation was then integrated into the MODIS scene using multiple linear regression. Incorporating LSU-derived fractions with 43 other metrics, we employed the random forest classifier based on an improved classification scheme to generate the land cover maps. The achieved overall accuracy of 91% surpassed previous studies, significantly reducing mapping uncertainties. Our analysis revealed that over the past two decades, 43.84% of NEB experienced land cover changes. Savanna and shrubland areas decreased by 23% (~167,000 km2) and 36% (~170,000 km2), respectively, in 2020, while grassland areas more than doubled, showing a 140% increase (~275,000 km2). This substantial increase warrants further investigation into potential links to pastureland expansion. Forests reduced by ~10,000 km2 (10%), with the decline being more pronounced in and around protected lands, evidencing strong pressure on these areas. Overall, our methodology demonstrates promising results for regional LCC analysis by utilizing multiple sensors to minimize uncertainties in mapping. This approach can effectively support policymaking and strengthen the science-policy interface, contributing to a more resilient and ecologically balanced future for the regions ecosystems and communities.
AreaSRE
Arrangementurlib.net > BDMCI > Fonds > Produção a partir de 2021 > CGCT > Unveiling Land Cover...
doc Directory Contentthere are no files
source Directory Contentthere are no files
agreement Directory Content
agreement.html 11/12/2023 13:49 1.0 KiB 
4. Conditions of access and use
Languageen
User Groupsimone
Visibilityshown
Update Permissionnot transferred
5. Allied materials
Next Higher Units8JMKD3MGPCW/46KUATE
Host Collectionurlib.net/www/2021/06.04.03.40
6. Notes
Empty Fieldsarchivingpolicy archivist callnumber copyholder copyright creatorhistory descriptionlevel dissemination doi e-mailaddress edition editor format isbn issn keywords label lineage mark mirrorrepository nextedition notes numberoffiles numberofvolumes orcid organization pages parameterlist parentrepositories previousedition previouslowerunit progress project publisheraddress readergroup readpermission rightsholder schedulinginformation secondarydate secondarymark serieseditor session shorttitle size sponsor subject targetfile tertiarymark tertiarytype type url volume
7. Description control
e-Mail (login)simone
update 


Close