Kevin O. Pope

Radar remote sensing of forest and wetland ecosystems in the Central American tropics

Abstract We analyzed airborne synthetic aperture radar (AIRSAR) imagery of forest, wetland, and agricultural ecosystems in northern Belize, Central America. Our analyses are based upon four biophysical indices derived from the fully polarimetric SAR data: the volume scattering index (VSI), canopy structure index (CSI), biomass index (BMI), calculated from the backscatter magnitude data, and the interaction type index (ITI), calculated from the backscatter phase data. We developed a four-level landscape hierarchy based upon clustering analyses of the 12 index parameters (four indices each for P, L, and C band) from two test site images. Statistical analyses were used to examine the relative importance of the 12 parameters for discriminating ecosystem characteristics at various landscape scales. We found that ITI was the most important index (primarily C band = CITI) for level, vegetated terrain at all levels of the hierarchy. BMI was most important for differentiating between vegetated and nonvegetated areas and between sloping and level terrain. These findings indicate that upper canopy spatial characteristics and flooding in marshlands (reflected in the CITI) are more important than biomass in differentiating many tropical ecosystems with radar data. The relative importance of the indices varied with vegetation type; for example, PVSI was the most important for distinguishing between upland forests and regrowth, and PCSI was the most important for differentiating swamp forest types. Finally, we evaluated the potential of present and future spaceborne SARs for tropical ecosystem studies based on our results. Most of these SARs are single channel systems and will provide limited capability for characterizing biomass and structure of tropical vegetation. This is especially true for C band systems, which produce data similar to our CBMI parameter, which was one of the least important in our analyses. The SIR-C/X-SAR and proposed EOS SAR are future spaceborne multifrequency fully polarimetric SAR systems, and they will provide a significant contribution to tropical ecosystem studies.

Detecting seasonal flooding cycles in marshes of the Yucatan Peninsula with SIR-C polarimetric radar imagery☆

Abstract Polarimetric L- and C-band radar imagery from the shuttle imaging radar-C (SIR-C) were acquired over wetlands of the Yucatan Peninsula during the dry (April) and wet (October) seasons of 1994. Field surveys during the flights recorded biophysical data and water depth in 11 marsh sites containing communities of three principal emergent macrophytes: Cladium jamaicense, Typha domingensis, and Eleocharis cellulosa. The only major seasonal change was in flooding. Seasonal changes in polarimetric backscatter magnitude (HH, VV, and CS=(HV+VH)/2) and phase [βH-V phase differenceβ=PD) were extracted for a stable evergreen mangrove forest calibration site, which confirmed that the absolute calibration of the Yucatan imagery exceeded the SIR-C system calibration. We estimate that seasonal changes of ⩾2dB in backscatter magnitude and ⩾10° in phase (PD) are significant in our data. Seasonal changes in L- and C-band magnitude and phase were extracted from the 11 marshes, and significant changes above the calibration limit were noted. Increased flooding in the marshes was detected by: 1) an increase in backscatter magnitude in marshes with tall, dense cover; 2) a decrease in backscatter magnitude in marshes with short, sparse cover, and 3) an increase in PD in all types of marshes. Magnitude increases result from an increase in double-bounce interactions between the emergent vegetation and water surface, whereas decreases result from an increase in forward scattering off the open water. Average PD values increase owing to an absolute or relative increase in double- compared with single-bounce interaction. Changes from dry or partially flooded to completely flooded, as well as increases in water depth, could be detected by most of the polarimetric parameters, but changes from dry to partially flooded could not. C-band PD (CPD) was the radar parameter most sensitive to flooding. CPD changed significantly for all eleven marshes, followed by L-band PD (LPD) and LVV (nine marshes) and LHH, LCS, and CVV (seven marshes). CHH detected significant changes in five marshes but produced changes of ±1.8–1.9 dB (just below our estimated calibration limit) in four others. An evaluation of current spaceborne radars indicates that a combination of the European Remote Sensing Satellite (ERS-1,2) and Radarsat radars could detect seasonal flooding in a wide variety of marsh ecosystems, excluding partial flooding and flooding in small patches of short, sparse vegetation.

Remote Sensing of Tropical Wetlands for Malaria Control in Chiapas, Mexico

Malaria, transmitted by anopheline mosquitoes, remains a serious health problem in the tropics. Most malaria eradication efforts focus on control of anopheline vectors. These efforts include the NASA Di-Mod project, whose current goal is to integrate remote sensing, geographic information systems (GIS), and field research to predict anopheline mosquito population dynamics in the Pacific coastal plain of Chiapas, Mexico. Field studies demonstrate that high larval production of Anopheles albimanus, the principal malaria vector in the plain, can be linked to a small number of larval habitat-types, determined by larval sampling and cluster analysis of wetlands in the coastal plain. Analysis of wet and dry season Landsat Thematic Mapper (TM) satellite imagery identified 16 land cover units within an 185-km2 study area in the coastal zone. A hierarchical approach was used to link the larval habitat-types with the larger land cover units and make predictions of potential and actual low, medium, and high anopheline production. The TM-based map and GIS techniques were then used to predict differences in anopheline production at two villages, La Victoria and Efrain Gutierrez. La Victoria was predicted to have much higher Anopheles albimanus production, based upon a 2-10 times greater extent of medium- and high-producing land cover units in its vicinity. This difference between villages was independently supported by sampling (with light traps) of adults, which were 5-10 times more abundant in La Victoria.

Landscape Ecology and Diversity Patterns in the Seasonal Tropics from Landsat TM Imagery

We examine patterns of landscape diversity in the seasonal tropical forests of northeastern Guatemala with analyses of Landsat Thematic Mapper (TM) imagery. Our primary objective is to demonstrate that TM analyses, without extensive field data, provide valuable information on landscape diversity patterns to aid conservation and development plans. A principal components analysis indicates that the major source of variability at the scale of TM resolution (=30 m) is related to TM band 4 (TM4) radiance, and that the second and third most important sources are related to TM bands 5 and 7, respectively. Cluster analysis of TM4, TM5, and TM7 radiance produced six distinct land-cover types, which follow a topographic gradient. Both TM radiance and the radiance coefficient of variation (CV) are significant in discriminating land-cover types. The primary source of variability within upland forests is related to TM4, and the primary sources of variability within lowland swamps are related to TM5 and TM7. We infer from these results that green leaf biomass is the most important variable in discriminating between land-cover types in the uplands, and that canopy closure and degree of senescence are the most important variables in the lowland swamps. We also examined patterns of landscape diversity reflected in three landscape indices: the number of land-cover types (LCT), the Shannon-Weaver index of landscape evenness (S-W), and a topographic index (TI). The strongest correlations were found for TM4 radiance and TI (r = 0.94, n = 30, P < 0.0001) and S-W (r = -0.72, n = 30, P < 0.0001), and between TM4 radiance CV and LCT (r, = 0.67, n = 30, P < 0.0001). Thus, TM4 provides a good measure of landscape diversity. We found a perfect nonparametric correlation (r, = 1, n = 6) between TM4 radiance CV within a land cover type and the types TI rank, which probably reflects increasing canopy heterogeneity be- tween lowland swamps and upland forests. Finally, we demonstrate the potential of TM- based land-cover maps and diversity indices in designing and monitoring nature reserves.

Distribution of Anopheles albimanus, Anopheles vestitipennis, and Anopheles crucians Associated with Land Use in Northern Belize

Abstract Anthropogenic land use changes often alter natural patterns of disease transmission. The goal of this study was to determine whether phosphorus input from sugarcane, Saccharum officinarum L., cultivation in northern Belize could pose a significant environmental impact on malaria transmission by changing vegetation structure and composition of wetlands and associated larval habitats. Our primary focus was on the increased dominance of cattail, Typha domingensis Pers., a favored habitat for Anopheles vestitipennis Dyar & Knab. A land cover classification based on satellite imagery was used to select 20 marshes impacted by agricultural runoff and 20 marshes surrounded by forest (nonimpacted). A 100-m transect was established into each of the 40 marshes. Water, vegetation, and larval sampling were conducted at the 0-, 10-, 25-, 50-, and 100-m locations along the transect. Analyses of larval density data indicated that Anopheles albimanus Wiedemann was negatively correlated with percentage of cover of Typha (R2 = 0.39, P < 0.001) but positively correlated with sparse Eleocharis cellulosa Torr. (rush) cover (R2 = 0.19, P < 0.05) and presence of cyanobacterial mats (CBM) (R2 = 0.33, P < 0.0001). An. vestitipennis was found to be positively correlated with percentage of cover of Typha (R2 = 0.19, P < 0.001). Canonical correspondence analysis identified CBM and light as the variables associated with the presence of An. albimanus larvae, Typha cover with An. vestitipennis larvae, and Eleocharis and absence of light with Anopheles crucians (Wiedemann). A positive correlation also existed between marshes adjacent to agricultural activities and presence of An. vestitipennis (R2 = 0.37, P < 0.05). These results indicate that marshes in proximity to agricultural fields are conducive for Typha growth, thereby providing habitat for the more efficient malaria vector An. vestitipennis.

Spaceborne imaging radar-C (SIR-C) observations of groundwater discharge and wetlands associated with the Chicxulub impact crater, northwestern Yucatan Peninsula, Mexico

Analyses of spaceborne imaging radar-C (SIR-C) data and field data from the northwestern Yucatan Peninsula, Mexico, demonstrate that spaceborne multifrequency polarimetric radars are excellent tools for characterizing patterns of wetland flooding. Seasonal flooding can be detected in most types of forest and marsh in the radar backscatter magnitude and phase data of both L and C band. Field observations made in the wet and dry seasons concurrent with the space missions and chemical analyses of floodwaters confirm that flooding is the product of discharge from the Yucatan aquifer, which consists of a fresh-water lens floating on seawater. This discharge controls the distribution of wetlands. Therefore, vegetation and flooding patterns, mapped with SIR-C imagery, provide valuable information on the hydrogeology of the region. Radar-image maps of wetlands and flooding indicate that there are three major zones of groundwater discharge that correlate with structures of the buried Chicxulub crater—zone 1 with the peak ring, zone 2 with the crater rim, and zone 3 with the exterior ring. Zone 1 has sulfate-poor discharge, unlike the sulfate-rich discharge in zones 2 and 3. The highest discharge is in zone 3, where the buried crater is closest to the surface. This groundwater-discharge pattern can be explained by tidal pumping of fresh water to the surface through high permeability zones developed in the Tertiary carbonates overlying crater faults and escarpments.

Overview of field studies for the application of remote sensing to the study of malaria transmission in Tapachula, Mexico

Abstract A National Aeronautics and Space Administration (NASA) sponsored project to use remote sensing technology in a predictive model of vector population dynamics and malaria transmission potential for the coastal plain of Chiapas, Mexico, is described. Included are the results of recent studies to characterize vector habitats and an assessment of the kinds of information that will be required for developing the predictive model within a geographic information system.