Study of the Floodflow Dynamics in the Pantanal of Cáceres/MT

Edinéia Aparecida dos Santos Galvanin

, Carla Bernadete Madureira Cruz

, Raúl Sanchez Vicens

Murilo Henrique Xavier Pereira

and Sandra Mara Alves da Silva Neves

Department of Mathematics, University of State of Mato Grosso, Barra do Bugres, Brazil

Department of Geography, Federal University of Rio de Janeiro, Rio de Janeiro, Brazil

Department of Geography, Fluminense Federal University, Niterói, Brazil

Department of Computing, University of State of Mato Grosso, Street A, s/n, Barra do Bugres, Brazil

Department of Geography, University of State of Mato Grosso, Cáceres, Brazil

Keywords: Land Cover, Geotechnologies, Wetland.

Abstract: The Brazilian Constitution lists some biomes and ecosystems considered as national patrimony, among which

the Pantanal Matogrossense is inserted, legally establishing conditions that ensure the preservation of the

environment, including the use of its natural resources (Article 225, § 4). In this context, the research proposes

to analyze the dynamics of the floodflow in the Pantanal of Cáceres/MT and aims to contribute a methodology

to obtain results that can support the planning, management, and monitoring of natural resources, as well as

providing important information for agriculture, geology, hydrology, and ecological models. In this research,

images from the MODIS sensor from the year 2015, with a resolution of 250 m, were used for the application

of the methodology. The software ArcGis and Matlab were used for the image processing. The normalized

difference water index obtained through 500 m resolution MODIS images was used as a priori analysis of the

regions with water and without water, after a threshold was defined for all of the images and the processing

was done to obtain the matrix with days with water and without water in the studied area. The results obtained

provide a classification for the floodflow, the quality parameters derived using study area show that the

proposed method performed better with classes < 30 (96%) and 180 - 270 (90%) of 4% and 10% of false

positives, respectively.

1 INTRODUCTION

The Pantanal is a large wetland composed of several

types of vegetation (landscape units) that make up a

complex set of habitats with multiple functions.

These systems depend on the pulse of flood and the

interaction of these environments. Biodiversity

maintenance needs these habitats. Similarly, this

immense floodplain maintains a relationship of

interdependence with the ecosystems located around

the Pantanal, considering, in this context, the impacts

caused by human intervention in the highlands

surrounding the wetlands (Wantzen et al., 2008; Silva

and Girard, 2004).

The Pantanal of Cáceres, as a subregion of the

Mato Grosso Pantanal, corresponds to an extensive

area of accumulation with a very flat topography that

is frequently subjected to floods, whose drainage

network is controlled by the Paraguai River (Brasil,

1997).

Some studies have concluded that the floods of

area are due to the volume of water brought by the

network of tributaries of the Paraguai River, together

with the weak slope and soil type rather than the

rainfall. The Pantanal vegetation consists of a variety

of species adapted to the dynamics defined by the

flood pulses.

According to Adamoli (2005), the Pantanal is a

biome that frequently undergoes alteration in its

vegetation cover. This factor is due to the

differentiated dynamics, caused by the double

seasonality and alterations in the flood regime, which

changes the structure and floral composition, making

the humid areas more sensitive to anthropic processes

(Bove et al., 2003).

Due to this distinct characteristic, flooded areas

require monitoring to verify the change in vegetation.

Santos et al., (2009) emphasize that the most efficient

way to promote the monitoring of these areas is

through Geographic Information System (GIS) and

Galvanin, E., Cruz, C., Vicens, R., Pereira, M. and Neves, S.

Study of the Floodﬂow Dynamics in the Pantanal of Cáceres/MT.

DOI: 10.5220/0006309001950200

In Proceedings of the 3rd International Conference on Geographical Information Systems Theory, Applications and Management (GISTAM 2017), pages 195-200

ISBN: 978-989-758-252-3

195

geoprocessing that makes it possible to have precise

evaluations, aiding in the analysis of the spatio-

temporal modifications. Due to the synoptic view,

mapping and repetitive coverage through remote

sensing images is a viable source of information at

regional and global scales (Csaplovics, 1998; Foody,

2002).

In this context, high-spatial revisits with a

moderate resolution image spectroradiometer

(MODIS) have been used to monitor land use/land

(LULC) cover, predict global changes, and to assist

in the protection of our environment. In Song et al.

(2011), an approach is proposed for the mapping of

LULC in the Amur river basin using MODIS at 250m

resolution, the normalized difference vegetation

index (NDVI), land surface vegetation index (LSWI),

and the reflectance time series data for 2001 and

2007.

Almeida et al. (2015) analyzed the spatio-

temporal variability of the Pantanal vegetation

cover by a principal component analysis applied to

a complete annual dataset of filtered EVI2 images.

PCA-based approach was able to capture the

essentials of the phenological/environmental

variability.

Gu et al. (2008) evaluated the relationship

between satellite derived vegetation indices

(normalized difference vegetation index–NDVI and

normalized difference water index–NDWI) and soil

moisture to understand how these indices respond to

soil moisture fluctuations.

The most relevant aspect of this work is the

exploration of methodologies to identify the

floodflow dynamics in the Pantanal of Cáceres, which

were formulated based on the analysis of the images

of the annual time series of the MODIS system with

250 m resolution and the normalized difference water

index (NDWI) of MODIS at 500 m resolution.

The scientific and technological relevance of the

theme, in the context of the development of new

methodologies for the efficient capture of

cartographic information, is evidenced by the

importance given to the theme by the International

Society for Photogrammetry and Remote Sensing

(ISPRS), where one of the terms of reference is the

remote sensing of land use and coverage.

This paper is organized into four sections: the next

section presents the proposed methodolgy, followed

by the experimental results, discussion, and the main

conclusions. Future prospectives are described in the

last section.

2 MATERIALS AND METHODS

2.1 Study Area

The Pantanal of Cáceres is one of the subregions of

the Mato Grosso Pantanal, which corresponds to

approximately 9.01% of its territorial area (Silva and

Abdon, 1998) and occupying 50.87% of the territorial

area of the municipality of Cáceres, located in Mato

Grosso. It is located in the upper Paraguai Basin

(BAP) in the southwest region of the state of Mato

Grosso (Figure 1).

It lies between the Paraguai river and the

municipality of Corumbá/MS (north–south) and

borders the Republic of Bolivia and the Poconé

Pantanal (east–west) at the geographic coordinates

15°31’15’’ and 17°37’45’’ South latitude and

58°32’30’’ and 57°21’55’’ West Longitude. The

Cáceres Pantanal area is 12,412.56 km

, where

12,371 km

(99.66%) are located within the

municipality of Cáceres (Neves, 2008).

Figure 1: Location map of Pantanal of Cáceres, Mato

Grosso, Brazil.

The study area represents an extensive flooded

environment, described as an ancient sandy alluvial

plain (Ab'Saber, 2006) with phyto-physiognomic

aspects composed of different vegetal types such as:

wooded savanna, forest savanna, and grass savanna

(IBGE, 2012).

The annual average temperature is 22.6°C (Brasil,

2007). It presents a rainfall index of 1200 to 1500 mm

GISTAM 2017 - 3rd International Conference on Geographical Information Systems Theory, Applications and Management

196

annually (Neves et al., 2011) with an altitude varying

between 90 and 200 meters (Radambrasil, 1982).

The soil of the region presents low fertility,

mainly composed by Plintossolo (PL), Plantossolo

(PT), and quartz sands (Quartzarenic Neosol), with

the influence of hydromorphic processes (Fernandes

et al., 2007; Embrapa, 2006).

2.2 Methodology

In this paper, we propose an analyze of the dynamics

of the floodflow in the Pantanal of Cáceres/MT.

Methodological steps are explained in detail in the

Figure 2.

Figure 2: Flowchart of methodology.

In the first stage of the methodology, 46 MODIS

product Q1 images were selected with a spatial

resolution of 250 m on the Terra platform, available

in the EOS Data Gateway (NASA, 2007) portal, in an

8-day composition, referring to the months of January

to December 2015.

The choice of the 8-day composition in this work

is due to the reduction of the number of images with

clouds and the effects of bidirectional reflectance and

atmosphere reflectance (Martinez et al., 2009; Villar

et al., 2012).

In this stage, the study area was cut by mask of the

area using the shapefile extension. The linear cut with

one standard deviation was used as a parameter for

the choice of threshold. The thresholds were chosen

in the software ArcGis, version 9.2 (ESRI, 2007).

The NDWI images of the MODIS product A1

with a spatial resolution of 500 m from the months of

January and July of 2015 was used to subsidize the

choice of threshold. The NDWI was calculated

according to Gao (1995).

Larger thresholds were tested, however, they

eventually underestimated the delimitation of the

areas with water, causing the grouping of areas with

differentiated responses. The use of one standard

deviation was an empirical choice but provided a

plausible result for determining the threshold to be

used in all study images.

It is worth noting that the months of the dry season

are May–October, with July being the month with the

lowest total rainfall. The months of the rainy season

are November–April, when 76% of the annual total

rainfall occurs, with January being the wettest month.

The resulting image was classified in a binary

image (with water values being 1 and not water being

0) and exported to Matlab software.

In the second step of the methodology, the images

were imported into the Matlab software, in which all

images were converted to a matrix, and to performed

the sum of the values of the pixels with water (value

of 1) and without water (value of 0) and the after the

product matrix was generated, resulting from the

matrix sum multiplied by the composition of the

MODIS images (8 days).

This gave a matrix with values of days with water

varying from 0 (not one day with water) to 365 (every

day with water). At the end of the compilation process

in the Matlab software, the georeferenced product

image and a classification of the days with water in

the study area were imported into ArcgGis.

For the classification, five different thematic

classes were defined, including the days with water:

<30 (nonflooded or rarely flooded––automatic

landscapes), 30–90 (occasionally flooded), 90–180

(seasonally flooded), 180–270 (frequently flooded),

>270 (permanently flooded–hydromorphic

landscapes).

Visits were made to the study area during the dry

period, in November 2015, to record photographs of

the existing landscapes in the region (attached in

Figure 1) and collect land control points (PCTs) to

subsidize the classification of satellite images.

After the classification process an assessment of

accuracy was performed using the Khat statistic to

check the reliability of the map generated (Congalton

(1991). The producer's accuracy refers to the total

number of correct pixels in a class divided by the total

number of pixels of that class as derived from the

Study of the Floodﬂow Dynamics in the Pantanal of Cáceres/MT

197

reference data (i.e., the column total). The user's

accuracy is the total number of correct pixels in a

class divided by the total number of pixels that were

classified in that class (i.e., the row total).

For the generation of the confusion matrix, the

classification generated in ArcGIS were converted to

the feature class, through the Raster to Polygon tool.

In the ArcGIS, 150 (hundred and fifty) points

were created randomly, distributed in the each class.

In the ArcMap application, the Esri Image Service

(World Imagery) was used for point validation.

3 RESULTS

In order to aid in the detection of variations in water

availability in the environment, the NDWI was used

(Figure 3) which, according to Gao (1995), is more

sensitive to variations because the values of

reflectances corresponding to the average infrared

region are used.

However, with a resolution of 500 m, this product

was used to subsidize the choice of a linear cut

threshold for the histogram of the MODIS images at

250 m, considering that the study was carried out on

a regional scale.

The NDWI of the months with the highest

(January) and lowest (July) total rainfall was analyzed

and verified by the histograms that in the month of

July (driest period), a greater distribution of the

values occurs. In the month of January (the rainy

season) it is possible to verify a higher concentration

of the values. The same behavior can be verified in

MODIS images at 250 m. This analysis assisted in the

final classification of the floodflow dynamics.

Figures (4) and (5) show the frequency histograms

of the MODIS images at 250 m from January and

July, respectively, and the linear cut with the

threshold of a standard deviation.

Figure 4: Image histogram of January 2015.

Figure 3: NDWI images. A) NDWI image of January 2015

and; B) NDWI image of July 2015.

GISTAM 2017 - 3rd International Conference on Geographical Information Systems Theory, Applications and Management

198

Figure 5: Image histogram of July 2015.

The classes that represent the floodflow in the

Pantanal of Cáceres / MT are presented in figure 6.

Figure 6: Classification of the floodflow in the Pantanal of

Cáceres / MT.

Table 1 shows the results of the quality parameters

obtained for the five thematic classes. Four main

columns are shown in Table 1: column 1 identifies the

thematic classes; column 2 shows the user’s accuracy

(number of false positives); column 3 shows the

producer’s accuracy (number of false negatives); and

column 4 shows the KHAT accuracy is a KHAT

statistic (an estimate of KAPPA).

The quality parameters derived using study area

show that the proposed method performed better with

classes < 30 (96%) and 180 - 270 (90%) of 4% and

10% of false positives, respectively. In the class of 30

- 90 was obtained (30% of false positives). Because

this area is occasionally flooded and this fact difficult

the classification.

Table 1: Quality parameters.

User’s

accuracy (%)

Producer’s

accuracy (%)

Khat

accuracy

< 30

0,82

30 - 90

90 - 180

180 - 270

> 270

In the table 1, class > 270 present 4% of false

negatives and class 90 – 180 presented 22% of false

negatives. Less than ideal results, in terms of false

negatives. In conclusion, the method performance for

this experiment can be considered satisfactory.

The result obtained in this work is essential for the

monitoring of the impacts caused by excess or lack of

water, as well as the definition of the environments of

the Pantanal of Cáceres.

4 CONCLUSIONS

The results generated in this work showed the areas

with and without water and resulted in defining the

thematic classes of the floodflow of the Pantanal of

Cáceres / MT, Brazil. For that, a methodology was

developed using an annual time series of images of

the MODIS Q1 and A1 sensor for the year 2015.

Results from this study indicate that the floodflow

can be mapped into thematic classes, according to

five flood distinctions.

To evaluate the proposed method, the analysis

were conducted involving sample of classified

classes. In general, the method showed a satisfactory

performance, few false positives occurred and few

false negatives were verified.

The main challenge of this work was to obtain a

single threshold for the entire series of images from

the year 2015.

As future work will be performed, some statistical

analyses will be performed to evaluate the results

obtained by the temporal profile of the floodflow.

ACKNOWLEDGEMENTS

The authors would like to thank the project

“Temporal analysis of land use to define scenarios of

natural landscape change by human interventions in

Study of the Floodﬂow Dynamics in the Pantanal of Cáceres/MT

199

the Pantanal of Cáceres / MT” financed by

FAPEMAT.

REFERENCES

Ab’Saber, A. N. Brasil, Paisagens de Exceção: o litoral e

o Pantanal Mato-Grossense: patrimônio básico.

Editora Ateliê. Cotia S.P., p.57-59, 2006.

Adamoli, J. Zoneamento ecológico do Pantanal baseado no

regime de inundações. In: Encontro Sobre

Sensoriamento Remoto Aplicado a Estudos no

Pantanal, 1., 1995, Corumbá. Anais. São José dos

Campos: Inpe, p.15-17, 2005.

Almeida, T. I. R.; Penatti, N. C.; Ferreira, L. G.; Arantes,

A. E.; Amaral, C. H. Principal component analysis

applied to a time series of MODIS images: the spatio-

temporal variability of the Pantanal wetland, Brazil.

Wetlands Ecology Management, vol. 23, p. 737-748.

2015.

Bove, C. P.; Gil, A. S. B.; Moreira, C. B.; Anjos, R. F. B.

Hidrófitas fanerogâmicas de ecossistemas aquáticos

temporários da planície costeira do Estado do Rio de

Janeiro, Brasil. Acta Bot. Bras., v.17, n.1, p. 119-135,

2003.

BRASIL. Ministério do Meio Ambiente. Biomas. Pantanal:

Fauna e Flora. 2007. Disponível em:<

http://www.mma.gov.br/biomas/pantanal#footer>.

Acesso em: jan. 2015.

BRASIL. Ministérios do Meio Ambiente, dos Recursos

Hidricos e da Amazônia Legal. Plano de Conservação

da Bacia do Alto Paraguai (Pantanal) – PCBAP.

Diagnóstico dos meios físicos e bióticos: meio biótico.

Brasília, DF: MMA, v. 2, t. 3. p. 3-68, 1997.

Congalton, R. G. A review of assessing the accuracy of

classifications of remotely sensed data. Remote Sensing

of Environment. n. 37, 35-46, 1991.

Csaplovics, E. High Resolution space imagery for regional

environmental monitoring — status quo and future

trends. International Archives of Photogrammetry and

Remote Sensing, 32(7), 211-216, 1998.

EMBRAPA. Centro Nacional de Pesquisa de Solos (Rio de

Janeiro, RJ). Sistema Brasileiro de Classificação de

solos. 2. ed. – Rio de Janeiro : EMBRAPA-SPI, 2006.

ESRI. ArcGIS Desktop: release 9.2. Redlands, CA:

Environmental Systems Research Institute, 2007.

Fernandes, F. A.; Fernandes, A. H. B. M.; Soares, M. T. S;

Pellegrin, L. A.; Lima, I. B. T. de. Atualização do mapa

de solos da planície pantaneira para o Sistema

Brasileiro de Classificação de Solos. Corumbá:

Embrapa Pantanal, 2007.

Foody, G. M. Status of Land Cover Classification Accuracy

Assessment, Remote Sensing of Environment, 80, 185-

201, 2002.

Gao, B.C. A normalized difference water index for remote

sensing of vegetation liquid water from space, in SPIE's

1995 Symposium on OE / Aerospace Sensing and Dual

Use Photonics, Vol. 2480, Orlando, FL, 1995.

Gu, Y.; Hunt E.; Wardlow, B.; Basara, J.; Brown, J. F.;

Verdin, J. P. Evaluation of MODIS NDVI and NDWI

for vegetation drought monitoring using Oklahoma

Mesonet soil moisture data. Geophysical Research

Letters, vol. 35, p. 1-5, 2008.

IBGE- Instituto Brasileiro de Geografia e Estatística.

Manual Técnico da Vegetação Brasileira. Ed. 2. 2012.

Martinez, J. M.; Guyot, J. L., Filizola, N., Sondag, F.

Increase in suspended sediment discharge of the

Amazon River assessed by monitoring network and

satellite data. Catena, v.79. p.257-264. 2009.

NASA. National Aeronautics and Space Administration.

Earth observing system data gateway. Disponível em:

<http:// edcimswww.cr.usgs.gov/pub/imswelcome/>.

Acesso em: 15 out. 2007.

Neves, S. M. A. S.; Cruz, C. B. M.; Neves, R. J.

Operacionalização de propostas de Zoneamento

Turistico do Pantanal de Cáceres/MT, com suporte nas

geotecnologias. Ciência Geográfica, v. 18, n., p. 76-86,

2008.

Neves, S. M. A. S.; Nunes, M. C. M.; Neves, R. J.

Caracterização das condições climáticas de

Cáceres/MT Brasil, no período de 1971 a 2009:

subsídios às atividades agropecuárias e turísticas

municipais. Boletim Goiano de Geografia, v. 31, n. 2,

p. 55-68, 2011.

Radam-Brasil. Projeto RADAMBRASIL: Levantamento de

Recursos Naturais Vol. 27 - Corumbá. Rio de Janeiro:

Ministério das Minas e Energia, 1982.

Santos, J. S.; Pereira, G.; Shimabukuru, Y. E.; Rudorff,

B. F. T. Identificação de áreas alagadas no Bioma

Pantanal - Brasil -utilizando dados multitemporais

TERRA/MODIS. GEOGRAFIA, Rio Claro, v. 34,

Número Especial, p. 745-755, dez., 2009.

Silva, J.S.V.; Abdon, M.M.; Delimitação do Pantanal

brasileiro e suas sub-regiões. Pesquisa Agropecuária

Brasileira, Brasília, v.33, Número Especial, p.1703-

1711, out.,1998.

Silva, C.J.; Girard. P New challenges in the management of

the Brazilian Pantanal and catchment area. Wetlands

Ecology and Management, 12: 553-561, 2004.

Song, K.; Wang, Z.; Liu, Q.; Liu, D.; Ermoshin, V. V.;

Ganzei, S. S.; Zhang, B.; Ren, C.; Zeng, L.; Du, J. Land

use/land cover (LULC) classification with MODIS time

series data and validation in the Amur River Basin.

Geography and Natural Resources, Vol. 32, Issue 1, p.

9–15, 2011.

Villar, R. E.; Martinez, J. M.; Guyot, J. L.; Fraizy, P.;

Armijos, E.; Crave, A.; Bazán, H.; Vauchel, P.; Lavado,

W. The integration of field measurements and satellite

observations to determine river solid loads in poorly

monitored basins. Journal of Hydrology, v.44, p.45-54.

2012.

Wantzen, K. M.; Cunha, C. N.; Junk, W. J.; Girard, P.;

Rossetto, O. C.; Penha, J. M.; Couto, E. G.; Becker, M.;

Priante, G.; Tomas, W. M.; Santos, S. A.; Marta, J.;

Domingos, I.; Sonoda, F.; Curvo, M.; Callil, C. T.

Towards a sustainable management concept for

ecosystem services of the Pantanal wetland.

Ecohydrology & Hydrobiology, 8: P. 115-138, 2008.

GISTAM 2017 - 3rd International Conference on Geographical Information Systems Theory, Applications and Management

200