Introduction

Data products from the Land Processes Distributed Active Archive Center (LP DAAC) are used in many different applications. They play an important role in modeling, help to detect changes to the landscape, and are a way to assess ecosystem variables. A few of those applications, published between July and September 2015, are highlighted below. 

Amphibian Habitat Loss in Dominican Republic

This study uses Moderate Resolution Imaging Spectroradiometer (MODIS) vegetation data to evaluate habitat change for four amphibian species in protected and unprotected areas of the Dominican Republic.

Science Objectives

In their paper "Habitat suitability and protection status of four species of amphibians in the Dominican Republic" published in Applied Geography, Sangermano and others (2015) set out to evaluate changes in the habitat of four amphibian species in protected areas of the Dominican Republic. Three of the species are endemic to the islandof Hispaniola and are considered vulnerable; the fourth species is native to the Dominican Republic and is considered a near threatened species. These species require specific habitats and are highly susceptible to habitat degradation. The major threats to these species are logging, charcoal production, and infrastructure development, all of which lead to forest loss.

Instruments Used

The authors use vegetation data from Terra MODIS from 2000 to 2011. They chose to use MODIS due to its daily acquisition and frequent revisit of an area and its ability to capture larger areas.

Major Findings

The authors identified the habitats based on in-situ data and previous research. The authors examined land cover data, the availability of energy and water based on monthly temperature and precipitation measurements, and vegetation cover NDVI data derived from the vegetation indices (MOD13Q1) for the years 2000 to 2011. Changes in the Normalized Difference Vegetation Index (NDVI) vegetation cover time series were calculated using the ManneKendall statistic to detect any declines in vegetation productivity in forest areas.

During the study period, the authors estimated that the forested habitat decreased in size between 7.7 and 9.3 percent. The authors also found that the Dominican Republic does not have abundant habitat areas for the four amphibian species, with only 13 to 27 percent of the country being considered suitable habitat. Using NDVI, the authors were able to estimate and compare the rates of change in both the protected and unprotected areas. The authors reported that comparing NDVI values is useful when a detailed land cover map is not available. They also believe using these resources can help identify areas that would benefit from a new protection status and can help manage existing protected areas.

References

Publication Reference

Sangermano, F., Bol, L., Galvis, P., Gullison, R., Hardner, J., and Ross, G., 2015, Habitat suitability and protection status of four species of amphibians in the Dominican Republic: Applied Geography, v. 63, p. 55–65. doi:10.1016/j.apgeog.2015.06.002

Image References

Granule ID: MOD13Q1.A2011161.h11v07.005.2011179145417

Vegetation Mapping of Amazonian Megafan

Advanced Spaceborne Thermal Emission and Reflection Radiometer (ASTER) data and decision tree classification are used to map diverse vegetation types across the Viruá megafan in northern Brazil.

Science Objectives

When viewing satellite imagery of the Amazonian wetlands in northern Brazil, several triangular or fan-shaped areas are apparent. These areas are called ‘megafans’ and are the result of a distributary drainage system active during the late Quaternary period (0.5 – 1.0 million years ago). As megafans are typically vast and inaccessible, remote sensing is a preferred method to collect data about these areas.

Instruments Used

In this study, described in the paper "Mapping vegetation in a late Quaternary landform of the Amazonian wetlands using object based image analysis and decision tree classification" published in International Journal of Remote Sensing, ASTER imagery is used as an input to distinguish distribution of land cover types of the Viruá megafan, an approximately 916 square kilometer (~354 square miles) region located in central Brazil.

Major Findings

The vegetation of the megafan is generally described as campinarana, which is a collection of several vegetation types, including grasslands, shrublands, herbaceous/woodlands and forest. Distinguishing the various campinarana types requires a fine-scale resolution; previous studies have achieved, at best, 46 to 77 percent accuracy and generally do not encompass an entire megafan. By using remote sensing data and combining the various classifications into a decision tree, the authors automated the identification of many land cover types over the entire Viruá megafan and achieved an overall accuracy of 87 percent.

Initially, a red-green-blue (RGB) combination of ASTER bands 3, 2, and 1 from data collected January 28, 2003 was used to create a mask distinguishing campinarana from the adjacent forest area. Normalized Difference Vegetation Index (NDVI) values calculated from ASTER bands 2 and 3 were key to differentiate grass and shrub campinarana from woody and forested campinarana. ASTER band 1 distinguishes campinarana from adjacent wetland areas as well as woody campinarana from forest campinarana. Additional data from Landsat 5 Thematic Mapper, synthetic aperture radar images, elevation data from the Shuttle Radar Topography Mission, and Advanced Land Observing Satellite/Phased Array type L-band Synthetic Aperture Radar are used to further classify areas.

References

Publication Reference

de Oliveira Cordeiro, C.L., and de Fátima Rossetti, D., 2015, Mapping vegetation in a late Quaternary landform of the Amazonian wetlands using object based image analysis and decision tree classification: International Journal of Remote Sensing, v. 36, no. 13, p. 3397–3422. doi:10.1080/01431161.2015.1060644

Image Reference

Granule IDs

• AST_07_00301282003143626_20151008124340_14512
• AST_07_00301282003143634_20151008124339_14514

Flooding and Vegetation in the Okavango Delta

MODIS and Landsat time series were used to track flood duration and link it to vegetation patterns in the Okavango Delta, Botswana.

Science Objectives

Murry-Hudson and others (2015), in their paper "Remote sensing-derived hydroperiod as a predictor of floodplain vegetation composition" published in Wetlands Ecology and Management, monitored the Okavango Delta wetland in Botswana to establish a history of flooding in the area and understand how it relates to potential ecological changes.

Instruments Used

The team used Terra’ MODIS Surface Reflectance product (MOD09Q1). The authors started monitoring the area with satellite imagery from 1989. They considered MODIS to be an excellent choice for its monitoring potential to generate near-real time flood extent maps in the Okavango Delta wetland system.

Major Findings

In the study, two imagery time series of flooded areas are created. The first uses Landsat data from 1989 to 2007. MODIS Surface Reflectance data are used to fill in the missing information from the years 1991, 2003, and 2004. The second time series consists of 89 8-day 250 meter surface reflectance images (MOD09Q1) images collected from 2000 to 2007, which were obtained from the LP DAAC Data Pool. The authors stacked these MODIS images to create a flood duration map of the Okavango delta for each water year. The MODIS data were validated with in-situ community vegetation data and Landsat data.

The authors found that recurring flooding in the area varies between 2,000 and 3,600 square kilometers (~772 and 1,390 square miles), with maximum flooding reaching 7,200 square kilometers (~2,780 square miles) in 2006. In the time series, it was evident that peak flooding occurred during the months of August and September. The authors were able to examine differences in types of floods and the duration of inundation in the Delta. They were also able to distinguish between low-amplitude short-duration flooding, which occurred in 2003, and a larger and longer flooding, as seen in the image to the left from 2004. Both of these floods caused parts of the delta to be inundated for the entire year and other areas of the delta that are typically dry were inundated for half of the year. The authors then used this data in combination with in-situ measurements, to identify what type of vegetation is in the areas and how they relate to frequencies and durations of floods. They found that the maps created from MODIS Surface Reflectance Band 1 (MOD09Q1) produced flood maps with 89 percent or greater accuracy, as compared to field data and Landsat imagery. The authors have initiated a long-term monitoring program at the Okavango Research Institute to compare flooding extents and to identify classes of floodplain vegetation using MODIS data.

References

Publication Reference

Murry-Hudson, M., Wolski, P., Cassidy, L., Brown, M., Thito, K., Kashe, K., and Mosimanyana, E., 2015, Remote sensing-derived hydroperiod as a predictor of floodplain vegetation composition: Wetlands Ecology and Management, v. 23, p. 603–616. doi:10.1007/s11273-014-9340-z

Image References

Granule IDs

• MOD09Q1.A2004249.h20v10.005.2007313190837
• MOD09Q1.A2004249.h20v11.005.2007313195105