This is a study done by a team of students from INDIAN INSTITUTE of TECHNOLOGY, Delhi. The study was aimed at detecting biological diversity on Earth, tracking and predicting changes over time. The study is based on the fact that most of the species require specific temperature range, specific precipitation, specific terrain , specific vegetation etc. to survive properly. Basically, most of the species require specific environment conditions for their survival and if the environment changes, their existence may be threatened . We have used ground based observations and satellite data to carry out this study. This site is to display our study and provide our code and other resources to anyone who wants to study about biodiversity.

We used ground observation data along with NASA satellite-based remote sensing data of environmental variables, such as temperature, precipitation, and vegetation index to predict habitable places for a particular species around the globe.

Procedure(Explaining using a species of coyote(Canis latrans Say)):

1. Took occurrence data of the coyote from 2001 to 2020 in USA from https://www.gbif.org/ . (Fig. 3)

2. For each occurrence, the environmental variable values(temperature, precipitation etc.) were assigned along with latitude and longitude using data from https://datasearch.globe.gov/ and https://neo.sci.gsfc.nasa.gov/view.php?datasetId=TRMM_3B43M . Here, for understanding the model more clearly, statistical analysis of one of the environmental factors (temperature) for that species is shown in detail.

3. First of all, histogram of temperature is made. So now, we know the fraction of total occurrence of species at any temperature (Fig. 4). Basically, now we have an approximate range of temperature which is suitable for this particular species.

4. Now using the satellite data of world temperature, the points whose temperature is in the range are marked. These points represent that the probability of finding particular species is relatively much higher than other unmarked points(Fig. 5).

5. After doing the same for all environmental variables, we take the intersection of marked points from all the maps(Fig. 6). (Here as an example, I have shown intersection of three factors viz. temperature, aerosol optical thickness and vegetation index).

6. Thus points marked on the new map created, represent the most probable habitable places for the coyote.

We can always take more and more environmental variables and optimize the model to get better prediction.

So we can use this model to predict biodiversity of any region and any future time, if we have enough data. We can always optimize every step to get better and better prediction.