Elsevier

Journal of Environmental Management

Volume 167, 1 February 2016, Pages 175-184
Journal of Environmental Management

Research article
Four decades of land-cover, land-use and hydroclimatology changes in the Itacaiúnas River watershed, southeastern Amazon

https://doi.org/10.1016/j.jenvman.2015.11.039Get rights and content

Highlights

  • The main LCLU change was pastureland replacing the forest.

  • Protected areas save the forest in the active pastureland frontier in Amazon.

  • The conversion of forest to pasturelands is main process of LCLU changes.

  • LCLU changes is responsible for hydroclimatology changes in Amazon.

Abstract

Long-term human-induced impacts have significantly changed the Amazonian landscape. The most dramatic land cover and land use (LCLU) changes began in the early 1970s with the establishment of the Trans-Amazon Highway and large government projects associated with the expansion of agricultural settlement and cattle ranching, which cleared significant tropical forest cover in the areas of new and accelerated human development. Taking the changes in the LCLU over the past four decades as a basis, this study aims to determine the consequences of land cover (forest and savanna) and land use (pasturelands, mining and urban) changes on the hydroclimatology of the Itacaiúnas River watershed area of the located in the southeastern Amazon region. We analyzed a multi-decadal Landsat dataset from 1973, 1984, 1994, 2004 and 2013 and a 40-yr time series of water discharge from the Itacaiúnas River, as well as air temperature and relative humidity data over this drainage area for the same period. We employed standard Landsat image processing techniques in conjunction with a geographic object-based image analysis and multi-resolution classification approach. With the goal of detecting possible long-term trends, non-parametric Mann–Kendall test was applied, based on a Sen slope estimator on a 40-yr annual PREC, TMED and RH time series, considering the spatial average of the entire watershed. In the 1970s, the region was entirely covered by forest (99%) and savanna (∼0.3%). Four decades later, only ∼48% of the tropical forest remains, while pasturelands occupy approximately 50% of the watershed area. Moreover, in protected areas, nearly 97% of the tropical forest remains conserved, while the forest cover of non-protected areas is quite fragmented and, consequently, unevenly distributed, covering an area of only 30%. Based on observational data analysis, there is evidence that the conversion of forest cover to extensive and homogeneous pasturelands was accompanied by systematic modifications to the hydroclimatology cycle of the Itacaiúnas watershed, thus highlighting drier environmental conditions due to a rise in the region's air temperature, a decrease in the relative humidity, and an increase in river discharge.

Introduction

In the Amazon region, the conversion of forest to pasturelands in the modern era began in the early 1970s with the construction of the Trans-Amazon Highway (Fearnside, 2005), which resulted in opening up the forested tropical region to human settlement and natural resource exploitation (Laurance et al., 2009). By 2006, almost 95% of all deforestation in the Brazilian Amazon occurred within 5.5 km of roadways or less than 1 km from navigable rivers (Barber et al., 2014). Hence, areas under pressure from human settlement were found primarily along official roads in the so-called “arc of deforestation”, which intersects agricultural, ecological and cultural areas along the eastern and southeastern of the Amazonian frontier.

The forest-clearing process is driven by a combination of factors, primarily infrastructure, financial incentives, and immigration (Aldrich et al., 2012). Deforestation is also influenced by the geographical position of the arc in relation to a climatological (moist tropical to tropical wet and dry climate) and ecological (tropical rainforest to Brazilian savannah) transition zone (Coe et al., 2013). Although cattle ranching remains the dominant form of land use in the Amazon Region, large-scale commercial agriculture, such as soy bean croplands, have fundamentally changed the landscape in the southeastern Amazon (Martinelli et al., 2010, Morton et al., 2006). These LCLU changes can have a number of consequences for freshwater ecosystems, including changes in the runoff characteristics (Sajikumar and Remya, 2015), changes in river discharge (Lima et al., 2014), changes in headwater stream temperatures (Macedo et al., 2013), degrading riparian areas (Deegan et al., 2011), and changes in the hydrological cycle as a whole (D'Almeida et al., 2007, Wohl et al., 2012). These alteration processes have been responsible for dramatic changes in the Amazon rivers watersheds over the last 40 years, causing among other things greater discharge and increased sediment flux in the channels of the main rivers (Coe et al., 2011, Costa et al., 2003, Latrubesse et al., 2009).

Mapping and monitoring the long-term and large-scale remote land cover and land use (LCLU) changes in the moist tropical regions has only been possible since the early 1970s with the launch of the first satellite of the Landsat series in 1972, known as the Earth Resource Technology Satellite - ERTS-1. This event began an era of Earth's surface monitoring by the “work horses” of space-borne optical data generation. Over the following decades, the moderate spatial resolution and global coverage images of the Landsat has allowed for the mapping, monitoring and assessment of LCLU changes at the local (Sonter et al., 2014), regional (Souza et al., 2013), and global scales (Hansen et al., 2013).

To understand the land cover (e.g., tropical forest) and land use (e.g., pastureland) conversion processes and their impacts on regional hydroclimatology1, a detailed spatial and multi-temporal scale analysis is required. Hence, we decided to investigate these processes in the Itacaiúnas River watershed, a basin within the context of the Amazon catchments, which encompasses the Carajás Mineral Province, one of the largest mining provinces worldwide. We also extended our investigation to several and geographically indigenous lands and environmental protected areas (ILPAs), which are characterized as being sheltered from deforestation activities (Fig. 1). This study aims to i) assess the LCLU change using the multi-decadal Landsat dataset from 1973 to 2013, associated with increasing deforestation and pastureland activities and ii) evaluate the long-term impacts of LCLU changes on the hydroclimatology of the Itacaiúnas River watershed. The outcomes of this study will be useful to inform better management of forest and savanna ecosystems either through the creation of new protected areas, reforestation of degraded areas or increasing awareness of politicians of changes in climate in response to human activities.

Section snippets

Study area

The present study was carried out in the Itacaiúnas River watershed that drains an area of approximately 41,300 km2 and is located approximately 600 km southward of the Equator line (Fig. 1). The area, confined by the geographical coordinates are 05°10′ to 07°15′ S latitude and 48°37′ to 51°25′ W longitude, was originally covered by the Amazon rainforest and experiences a typical monsoon climate (Alvares et al., 2013). It experiences well defined rainy (November to May) and dry (June to

Geographic object-based image analysis (GEOBIA)

The proposed classification approach is based on GEOBIA (Blaschke, 2010). This approach included multi-date segmentation, a classification process based on segmentation with the extraction of object multi-date spectral signatures, multi-resolution classification using a rule set-based approach, and quantifying classification.

Segmentation is the process of partitioning an image into groups of pixels that have similar numerical characteristics and are spatially adjacent, thereby minimizing the

LCLU dynamic across the itacaiúnas river watershed

Deforestation in the Itacaiúnas River watershed significantly changed the LCLU over the period 1973 to 2013 in the studied area, as shown by Fig. 3. In 1973, a small area of the natural land cover was used for pastureland (32,756 ha), which corresponds to only 0.79% of the total studied watershed. However, in 1984, 1994, 2004 and 2013, pastureland occupied an area of approximately 10%, 28%, 46% and 50%, respectively, of the entire watershed, and currently, pastureland covers approximately 2

Discussion

Our results from GEOBIA showed that approximately 50% of the Itacaiúnas watershed remains covered with forest and a similar area was converted to pastureland. However, remaining area covered by forest, that is, only the area situated in the ILPAs domains, which corresponds to ∼33% of the total area should remain effectively unchanged. Similar rates of deforestation have also been observed along the arc of deforestation (Fearnside, 2005), where the opening of major roads in the early 1970s, such

Conclusions

We conclude that over the last four decades, the Itacaiúnas River watershed was subject to two different settings of LCLU changes: one marked by a successful model of a public-private partnership between Brazilian environmental agencies and a multinational mining company, leading to forest conservation inside ILPAs, and the other associated with a strong deforestation process and the formation of a new landscape dominated by pastureland. The deforestation reached 52% of the watershed area, and

Acknowledgments

The authors would like to thank the United States Geological Survey, Department of the Interior (USGS), for providing the Landsat-8 OLI and Landsat-5 TM images and the National (Brazilian) Institute for Space Research (INPE) for providing Landsat-1 MSS data. The authors thank the members of DIPF, GELIF, DIST, LISF, LAMSF and GABAN of Vale S.A for the field support. This project was partially carried out in the National Forest of Carajás with the permission of IBAMA (SISBIO 35594-2). The first

References (51)

  • W.F. Laurance et al.

    Impacts of roads and linear clearings on tropical forests

    Trends Ecol. Evol.

    (2009)
  • L.A. Martinelli et al.

    Agriculture in Brazil: impacts, costs, and opportunities for a sustainable future

    Curr. Opin. Environ. Sustain.

    (2010)
  • N. Sajikumar et al.

    Impact of land cover and land use change on runoff characteristics

    J. Environ. Manage

    (2015)
  • L.J. Sonter et al.

    Processes of land use change in mining regions

    J. Clean. Prod.

    (2014)
  • S. Aldrich et al.

    Contentious land change in the Amazon's arc of deforestation

    Ann. Assoc. Am. Geogr.

    (2012)
  • C.A. Alvares et al.

    Koppen's climate classification map for Brazil

    Meteorol. Z.

    (2013)
  • M. Baatz et al.

    Multiresolution segmentation: an optimization approach for high quality multi-scale image segmentation

  • P.M. Brando et al.

    Ecology, economy and management of an agroindustrial frontier landscape in the southeast Amazon

    Philos. T. Roy. Soc. B

    (2013)
  • N.A. Chappell et al.

    Identifying step changes in single streamflow and evaporation records due to forest cover change

    Hydrol. Process

    (2012)
  • M. Chen et al.

    Assessing objective techniques for gauge-based analyses of global daily precipitation

    J. Geophys. Res.

    (2008)
  • M.T. Coe et al.

    The effects of deforestation and climate variability on the streamflow of the Araguaia River, Brazil

    Biogeochemistry

    (2011)
  • M.T. Coe et al.

    Deforestation and climate feedbacks threaten the ecological integrity of south–southeastern Amazonia

    Philos. T. Roy. Soc. B

    (2013)
  • R.G. Congalton et al.

    Assessing the Accuracy of Remotely Sensed Data : Principles and Practices

    (2009)
  • C. D'Almeida et al.

    The effects of deforestation on the hydrological cycle in Amazonia: a review on scale and resolution

    Int. J. Climatol.

    (2007)
  • E.B. De Souza et al.

    Precipitação climatológica sobre a Amazônia oriental durante o período chuvoso: observações e simulações regionais com o RegCM3

    Rev. Bras. Meteorol.

    (2009)
  • Cited by (0)

    View full text