Banca de QUALIFICAÇÃO: MATHEUS LEAL SOARES DAS NEVES

Uma banca de QUALIFICAÇÃO de DOUTORADO foi cadastrada pelo programa.
STUDENT : MATHEUS LEAL SOARES DAS NEVES
DATE: 23/03/2026
TIME: 09:00
LOCAL: Sala 24 do departamento de Solos
TITLE:

Use of Remote and Proximal Sensing for Digital Mapping and Soil Moisture Monitoring in Southern and South-Central Rio de Janeiro State

KEY WORDS:

Soil moisture; remote sensing; spatial modeling; downscaling; autorA.

PAGES: 15
BIG AREA: Ciências Agrárias
AREA: Agronomia
SUMMARY:

Soil moisture is a key variable for applications in agriculture, hydrology, ecology, and environmental risk management, but its measurement and continuous spatial representation still constitute a challenge in Brazil, especially in regions with low density of in situ monitoring. In this context, this thesis proposes to develop a methodological basis for modeling, mapping, and monitoring soil moisture in the South and South-Central Fluminense regions, integrating remote sensing and proximal data in a multiscale approach. The central hypothesis argues that the harmonization of multi-source data, associated with sampling optimization techniques and spatio-temporal modeling, allows for the precise and accurate representation of soil moisture dynamics and expands the spatial extrapolation capacity of the models.

The methodology is organized into four complementary fronts. The first consists of delimiting reference areas using the autorA algorithm, fed by environmental covariates such as relief, climate, geology, soil, and vegetation indices, in order to identify environmentally similar portions of the landscape and guide a more efficient sampling plan. The second front involves the temporal synchronization between field campaigns and SMAP satellite passes, seeking to make in situ measurements compatible with the ~9 km orbital product. The third comprises the transfer of predictive models developed in external and similar areas to the area of interest in Brazil, followed by validation with data obtained in the field by proximal sensors and by gravimetric determination. The fourth front corresponds to the spatial refinement of the estimates through downscaling techniques, using higher resolution images, such as those from Sentinel-2.
The thesis's sampling design was structured in two integrated stages. The first includes the local validation of the SMAP product in a pilot area in the municipalities of Seropédica, Itaguaí, and Paracambi, with evaluation of different sampling densities per pixel. The second expands the plan to the entire area of interest, in the South and South-Central Fluminense regions, with sensitivity analysis of the number of sampling points. Preliminary results indicate that intermediate scenarios offer a better balance between predictive performance and operational feasibility, with the n_130 set standing out, with an R² of 0.83, RMSE of 3.65, and RPIQ of 3.37.
It is expected that, in the end, a robust, replicable, and more economically efficient methodological protocol will be consolidated for the validation, modeling, and spatial refinement of soil moisture in tropical environments. As a contribution, the research may support monitoring networks, assist in agricultural management, strengthen hydrological analyses, and provide technical support for territorial and environmental decision-making.

 

COMMITTEE MEMBERS:
Externo à Instituição - DIEGO NUNES BRANDÃO - CEFET/RJ
Externo ao Programa - 3442355 - ERLI PINTO DOS SANTOS - nullExterno à Instituição - GUSTAVO DE MATTOS VASQUES - EMBRAPA
Presidente - 1220296 - MARCOS BACIS CEDDIA