NASA Marshall Space Flight Center
North Mexico Ecological Forecasting
Days of Our Ocelot: Finding an Elusive Cat's Habitat in North Mexico 

[music]


>> Maggi: The ocelot is a medium-sized cat that can be found from Argentina to the southern tip of Eastern Texas. Unfortunately, these cats are listed as endangered in the United States where there is reportedly less than 100 left. This decline in population is mainly due to habitat loss caused by deforestation, increased urbanization, road impacts, and illegal fur trade. Inbreeding is also an issue since the ocelot populations in Mexico and the United States are separated. As mentioned in an interview with one of our partners, Dr. Arturo Caso, aerial data is not currently used for monitoring ocelots in Mexico, and what imagery they do have is inaccurate. Tracking ocelots is expensive and time consuming requiring the usage of cameras or capturing ocelots and fitting them with radio telemetry or GPS collars. The objective of this project was to use remotely sensed data to create past, present, and future Habitat Probability Maps and a Habitat Percent Cover Graph to assess the extent of the ocelot habitat and to show which areas are most likely inhabited by ocelots. The area studied was northeastern Mexico, specifically Tamaulipas and parts of Nuevo Leon and San Louis Potosi. Data from 1996 to 2014 were used.

>> Ryan: This project used NASA's Landsat 5 TM and Landsat 8 OLI satellites to perform land classifications throughout the study area. These classifications were useful in determining which areas are preferred by the ocelots. Digital Elevation Models were obtained from SRTM version 2. Surface reflectance imagery from Suomi NPP VIIRS was used to derive a Normalized Difference Vegetation Index, or NDVI, which helped verify the land classifications.

>> Padraic: Using distance to streams, elevation, slope, land cover classifications,  and 19 bioclimatic variables, a Habitat Probability Map was created. The variables were then run in the Princeton Maximum Entropy model for the years 1996, 2004, and 2014. This determined the potential distribution of ocelots within the study area. Finally, the program was used to create future predictions of ocelot habitat probability for the years 2050 and 2070.

>> Leigh: 3 variables were derived from satellite data to determine which areas are likely inhabited by ocelots. Each variable was rescaled and combined to show a final Habitat Probability Map. This methodology was done for the years 1996, 2004, and 2014.

>> Padraic: The results of the Maximum Entropy Model showed a habitat decrease of over 170 square kilometers between 1996 and 2014. The amount of total habitat in the study area that is suitable for ocelots is less than 0.04%.

>> Leigh: The results of the Fuzzy Membership Model showed a habitat decrease from 1996 to 2004 and then a habitat increase from 2004 to 2014. In 2014, the suitable habitat made up more than 0.6% of the study area.

>> Padraic: Future projections predicted large gains by 2050. These gains are predicted to be eroded by 2070, but still represent an increase of over 78,000 square kilometers.  Even with those gains, suitable ocelot habitat is less than 0.04% of total within the study area.

>> Ryan: While every attempt was made to find and use data without clouds, some cloud cover in the imagery was unavoidable.  These clouds lead to increased errors in the land cover classifications. Repetitious data of erroneous data present in the Landsat 5 imagery also affected the accuracy of the land classifications.

>> Leigh: The team helped project partners better approximate ocelot population, current amount of suitable habitat, and a future projection of suitable habitat.  This knowledge provided support for conservation and research efforts.

[music]