Resilient Seeds

Gloria Otieno, Jeske van de Gevel, Maarten van Zonneveld

Learning objectives

At the end ofthis module, you will be able to:

• Download climate datasets and data on crop presence
• Prepare and import your dataset for use in DIVA-GIS, MaxEnt, and Google Earth
• Create online maps with the Climate Analogues tool 

The first step in identifying appropriate germplasm for adaptation measures isto gain anoverview of which geographic areas are suitable for a specific crop or variety. By combining biodiversity data (e.g., crop presence data for accessions held in national or international institutes) with climate data (e.g., the 19 bioclimatic variables in the WorldClim package), we will be able tomap this data for climate change analysis and germplasm identification. 

Sources of climate data

Climate data can be obtained from various sources, including local field observations, national meteorological stations, and global agencies and databases.

Regional meteorological data can be acquired by contacting individual regional and national weather stations. These data are usually observations of parameters, such as temperature and precipitation for a specific day and time. For this information tobe useful in detecting trends, it should span 30 years or more. However, in developing countries, weather stations are often few and do not cover large areas. Data may lack precision, and meteorological organizations might charge a fee for use of their data. The quality of the datasets and the format in which they are distributed depends on the organization.

Environmental sensors can be used to collect your own local weather information. For example, iButtons are small, affordable devices that measure temperature and humidity atpredetermined intervals. They can be placed in less accessible areas where local weather data are not available at a high resolution. Bioversity scientists have written a manual that describes in detail how to use iButtons for weather observation (Mittra et al. 2013). iButtons are bought in combination with devices to connect them to your computer and specialized software to process and store the readings in comma-separated variable (CSV) format.

Global agronomic weather data can be obtained via the aWhere platform, which combines observational weather data, forecast models, and historical norms from meteorological stations around the world and interpolates these into 9-km grid cells. Data include precipitation, minimum and maximum temperature, minimum and maximum relative humidity, solar radiation, maximum morning wind speed, and calculations such ascumulative growing degree-days (where you set base and cap temperatures).

aWhere’s data can be accessed freely by users in East and Southern Africa and parts of West Africa and India; however, you must register tobe able to view or download data via the aWhere platform (apps.awhere.com). aWhere also offers the possibility of uploading your own trial and weather data.

WorldClim is a set of global climate layers that can be downloaded for free via http://www.worldclim.org  [1]. These layers cover all global land areas except Antarctica andconsist of19 bioclimatic variables derived from monthly temperature and precipitation values. Bioclimatic variables represent annual trends, seasonality, and extremes. They are inthe latitude/longitude coordinate reference system (not projected), and the datum isWGS84. Table 1 summarizes various climate datasets.

Table 1: WorldClim and National Oceanic and Atmospheric Administration climate data

Datasets	Period covered	Source data	Available spatial resolutions
Present conditions Download via http://www.worldclim.org/current  [2]	1950–2000	Interpolations ofobservations	30 seconds 2.5 minutes 5 minutes 10 minutes
Projected future conditions Download via http://www.worldclim.org/CMIP5  [3]	To about 2050	Climate projections from the Fifth Assessment Report ofthe Intergovernmental Panel on Climate Change	30 seconds 2.5 minutes 5 minutes 10 minutes
Past conditions Download via http://www.worldclim.org/paleo-climate  [4]	Last glacial period (22 000 years ago) to mid-Holocene (6000 years ago)	Downscaled paleoclimate data	30 seconds 2.5 minutes 5 minutes 10 minutes
Surface marine data Download via http://icoads.noaa.gov/  [5]	1800 to present 1960 to present (more detailed)	Gridded monthly summaries using observations from many systems	2° × 2° longitude boxes (from 1800) and 1° × 1° boxes (from 1960)

The datasets are available at four spatial resolutions: from 30 seconds (0.93 km × 0.93 km = 0.86 km2 at the equator) to 2.5, 5, and 10 minutes (18.6 km × 18.6 km = 346.0km2at the equator). The highest resolution (30 seconds) is available per tile (region) in georeferenced TIFF format (GeoTIFF file) and can be used in any GIS application. WorldClim data at lower resolutions can be downloadedin ZIP-files in a generic grid (raster) format and in Environmental Systems Research Institute (ESRI) grids (for use with ESRI products). Table 2 summarizes the approximate grid resolutions.

Table 2: Approximate grid resolutions

Resolution	Approximate unit area
30 arc seconds	1 km²
2.5 arc minutes	5 km²
5 arc minutes	9 km²
10 arc minutes	18 km²

To import the generic grid (raster) format into DIVA-GIS, first download and unzip the ZIP-files. Use: Data\Import to gridfile\Multiple Files (BIL/BIP/BSQ). Some of these files have the extension .GRD. In this case, if you rename the .BIL files to .GRI, they can be opened inDIVA-GIS without the import procedure.

High resolution physical climate data can also be obtained from the Intergovernmental Panel on Climate Change (IPCC) through its data distribution centre. Climate model data from IPCC AR4 (2007) and IPCC TAR (2001) are viewed through the DDC file navigator and can be downloaded in CSV format. Data are available in10-year or30-year intervals ranging from 1900 to 2000 for the following variables: temperature, wet days, precipitation, daily maximum temperature, daily minimum temperature, ground frost frequency, water vapour, diurnal temperature range, and cloud cover.

Sources of biodiversity data

Biodiversity and environmental data can also be obtained from various sources, including (national) gene banks, herbariums, field observations, and global agencies and databases. A few of these sources are mentioned here.

Global or national gene banks serve as repositories of plant genetic materials. Passport and evaluation data for accessions can be obtained by contacting individual banks. Passport data include the source and origin ofan accession as well as taxonomic identification. Gene bank collections can also contain characterization and evaluation data on phenotypic traitsor provide more specific trait data including those of underutilized crops which are not covered in global databases. Examples of gene banks with clear trait data or the ability torequest germplasm online are listedin Table 3.

Table 3: Examples of gene banks that offer easily accessible and clear trait data

Name	Source
National Plant Germplasm System (also GRIN Global)	United States Department of Agriculture http://www.ars-grin.gov/npgs/orders.html  [6]
Centre for Genetic Resources (Dutch national gene bank)	Wageningen University and Research Centre, the Netherlands http://www.wageningenur.nl/en/Expertise-Services/Statutory-research-tasks/Centre-for-Genetic-Resources-the-Netherlands-1.htm  [7]
European catalogue ofex situ collections of plant genetic resources (EURISCO)	Secretariat of the European Cooperative Programme for Plant Genetic Resources http://eurisco.ipk-gatersleben.de/apex/f?p=103:1  [8]

Not all gene banks make their collection data available online. Quality and access to the data depend on the regulations and resources of individual institutes.

What software you use for data analysis will depend on the type and format of the data. Georeferenced climate and accessions data can be used to create multilayered maps in a wide range of GIS software programs. However, itis important to choose software that can do both mapping and analysis from points to grids to landscape. A number of free software programs exist for this type of analysis.

Species distribution models and crop models are used to predict climate change impact oncrop suitability and yield. Application of processed-based models may require specialized knowledge and a detailed set of parameters for specific areas. Therefore, niche models that predict climate change impact on crop suitability maybe a good option to provide recommendations in the context of the uncertainties involved in climate change projects and the often limited amount and quality of data.

Ecocrop

Ecocrop is a simple niche-based empirical model that uses environmental ranges to define the suitable area of a crop. It draws on the Food and Agriculture Organization’s Ecocrop plant database, which includes optimal environmental ranges of more than 2000 species. The model allows adjustment of minimum and maximum temperature based on local research findings. The Ecocrop database is available at www.ecocrop.fao.org  [15] and is also included in DIVA-GIS. Selecting a crop and setting parameters based on your own research findings allows you to create maps that show the suitability of a certain crop now and in the future.

Other resources for species distribution modeling:

• ModEco — integrated software for species distribution analysis and modeling
• DesktopGarp — software package for biodiversity and ecological research
• openModeller — a generic approach to species' potential distribution modelling

Mapping Software

In this module, we focus on four programs: DIVA-GIS to map the distribution of biological diversity and query climate data; MaxEnt, species distribution modeling software, to model the range in which a species can occur; Google Earth to create maps with georeferenced occurrence data or accessions data against high resolution satellite imagery; and Climate Analogues to project future climate conditions for a particular location and match this to sites that currently have similar rainfall and climatic conditions.

DIVA-GIS

DIVA-GIS is an open-source software program used to create maps and carry out geographic data analysis. It can create a wide range of maps, from a map of the world, to a map of a very small area, such as a district or even a village, showing, for example, state boundaries, rivers, a satellite image, and sites where an animal species was observed. This program can be downloaded at http://www.diva-gis.org  [17]

DIVA-GIS comes with the option to download free spatial data, such as administrative boundaries, roads, etc., and species occurrence data from GBIF, Genesys, LandSat, etc. As noted above, the Ecocrop model is built in.

Using DIVA-GIS, you can also download free spatial data  [18]  at http://www.diva-gis.org/Data  [18] for the whole world, which can then be used in DIVA-GIS or other GIS programs, such as Arc-GIS. DIVA-GIS is particularly useful for mapping and analyzing biodiversity data, such as the distribution of a species or other “point distributions.”

Climate data from WorldClim can be downloaded directly into DIVA-GIS at http://www.diva-gis.org/climate  [19]. This makes it possible to overlay climate information with species occurrence or other georeferenced data to provide an overview of the way the distribution of a species changes due to climate over certain periods. 

MaxEnt

Maximum entropy modeling uses layers of environmental variables (elevation, precipitation, etc.) as well as a set of georeferenced occurrence locations to produce a model of the range of a given species. One of the main applications of MaxEnt is prediction of species occurrence. From current species locations and environmental predictors (e.g., precipitation, temperature) across a user-defined landscape divided into grid cells, MaxEnt extracts a sample of background locations that it contrasts with present locations to predict species occurrence.

Available from: http://www.cs.princeton.edu/~schapire/MaxEnt/  [20]

Google Earth

Google Earth is a geobrowser that provides satellite and aerial imagery, ocean bathymetry, and other geographic data over the Internet to display the Earth as a three-dimensional globe. It has many features, including an increasing set of layers of mappable data, the ability to display third-party data, tools for creating new data, and the ability to import GPS data.

Georeferenced occurrence data or accessions data can be imported and mapped in Google Earth. The resolution is as high as 1 m and, therefore, specific names and locations can easily be obtained. The free version of Google Earth is downloadable at http://www.google.com/earth/download/ge/agree.html  [24]

Google Earth is searchable and allows the user to pan, zoom, rotate, and tilt the view of the Earth. Its layers of data, such as volcanoes and terrain, reside on Google’s servers, and can be displayed. Its elevation data, primarily from NASA’s Shuttle Radar Topography Mission,   [25]provide a terrain layer that adds depth to the landscape. Google Earth can also be used to acquire the coordinates of collections or occurrences with location names but no GPS information. 

Climate Analogues

Climate Analogues is an open-access tool developed by the programme on Climate Change, Agriculture and Food Security in conjunction with the International Center for Tropical Agriculture and the Walker Institute. Used to support adaptation to climate change in the agricultural section, its main applications are in agricultural policy and planning. The tool can be used to identify future climate conditions at a particular location, sites that currently resemble these conditions, and locations that have or will have similar climate conditions.

The tool can facilitate knowledge-sharing among communities, providing the opportunity to transfer practices and technologies to improve adaptive capacities. It can also provide insights into whether successful adaptation options in one location can be transferred to a future climatic analogue site. The temporal analogues are time specific and make use of past climates to create a representative time series for future climates. This allows the identification of historic events that might provide insight into the possible future consequences of climate change.

Based on careful analyses using the tool and supported by data from actual conditions in farmers’ fields, scientists can formulate possible intervention strategies, including identification of appropriate plant genetic resources, or develop new varieties for specific locations of interest.

The Climate Analogues online platform can be accessed at http://www.ccafs-analogues.org/tool/  [27]. The procedure for using this tool and interpreting analogue maps is described in a tutorial: http://www.ccafs-analogues.org/tutorial/  [28].

 

A common feature of the data referred to in this module is that they include georeferenced information and/or basic passport data. The data must be organized in a format that can be recognized by software such as DIVA-GIS and MaxEnt. These software applications have the capability of processing vector and raster data. Vector data come in the form of shapefiles with extensions, such as .SHP, .SHX, and .DBF, which store spatial features. Environmental data from specific geographic areas may be organized in rasters, which consist of a matrix of cells (or pixels) organized into rows and columns (or a grid) where each cell contains a value representing information, such as temperature. Rasters can be digital aerial photographs, imagery from satellites, digital pictures, or even scanned maps.
The level of detail in a raster is referred to as resolution. Raster sizes range from 1° (111 km) to 30 seconds (approximately 1 km at the equator).

Preparing data for use in DIVA-GIS and MaxEnt

1. Converting data into appropriate formats 

Presence points, which consist of passport data, can be entered or downloaded intoanExcel file and then converted into appropriate formats for spatial analysis using GIS applications, such as DIVA-GIS or MaxEnt. Your data should include an identification code, a scientific or taxonomic name, and coordinates (latitude and longitude). Other relevant information can also be added. Records downloaded from GBIFor Genesys will contain this information, but always check to ensure that the data are complete (e.g., no blank fields) and that the coordinates are in decimal degrees.

Table 4: Convert geographic coordinates from degrees, minutes, and seconds or degrees and decimal minutes to decimal degrees.

Decimal Degrees = [(Degrees (°) + Minutes (') / 60 + Seconds('') / 3600)] * H

H = 1 when the coordinate is  in the Eastern (E) or Nothern (N) Hemisphere
H = -1 when the coordinate is in the Western (W) or Southern (S) Hemisphere

Longitude	Degrees, Minutes & Seconds	Decimal Degrees	Latitude	Degrees, Minutes & Seconds	Decimal Degrees
Eastern Hemisphere	60°20'15''E	+ 60.3375	Northern Hemisphere	24°00'45''N	+ 24.0125
Western Hemisphere	60°20'15''W	- 60.3375	Southern Hemisphere	24°00'45''W	- 24.0125

Source: Scheldeman and van Zonneveld (2010: 24–26); see section on Google Earth above for full reference.

If the dataset you are working with is missing coordinates, you can add them manually using a gazetteer database. A gazetteer is an alphabetical database of administrative units combined with geographic coordinates. Gazetteer databases can be downloaded from the DIVA-GIS website (http://www.diva-gis.org/gdata  [30]).

Find a description of the location of your data point in your dataset; this could be the name of a village or town or a larger administrative unit, such as a municipality or a region. Use this information to search the gazetteer database for matches. Caution must be exercised to avoid assigning coordinates of a different village with the same name in another administrative unit. Alternatively, use Google Earth to locate the missing coordinates by typing in the description of the location and pinpointing the data point.
The placemark will contain the coordinates, which you can copy to your dataset.

In summary:

• Your dataset contains the required ID, label (taxonomic name), and latitude, longitude columns.
• Coordinates are in decimal degrees.
• Presence points with missing coordinates are removed completely from the dataset or missing coordinates are assigned using a gazetteer database.

2. Importing data into DIVA-GIS

Before continuing, make sure you have DIVA-GIS version 7.5 installed. Go to the DIVA-GIS website and click on the download page (http://www.diva-gis.org/download  [31]). For full functionality, you should also download the WorldClim climate data from the DIVA-GIS website (http://www.diva-gis.org/climate or from the downscaled GCM data portal (http://gisweb. ciat.cgiar.org/GCMPage  [32]).

Once you have DIVA-GIS up and running, take some time to familiarize yourself with the program. A good starting point is the DIVA-GIS 7.5 manual by Hijmans et al. (2012) which can be found at http://www.diva-gis.org/docs/DIVA-GIS_manual_7.pdf  [16]

Importing occurrence data: From the Data menu select “Import Points to Shapefile” and choose which type of file you want to import. DIVA-GIS allows direct imports from text (.TXT), Access database (.MDB), Excel (.XLS), or dBase (.DBF) files. You are asked to specify your dataset (input file) and the columns that contain latitude and longitude. Click “Save to Shapefile” to generate a vector file (.SHP). This shapefile contains all the presence points in your dataset.

Importing climate data (WorldClim): Previously, you downloaded climate layers from the DIVA-GIS website. The files in the ZIP-file are CLM files, which you should extract and store in a folder on your hard drive. To load the climate data into DIVA-GIS, go to Tools/Options/Climate and select the folder in which you stored the WorldClim files. Check to ensure that DIVA-GIS selected the right columns for each parameter and press “Apply.” You have now loaded the climate data into DIVA-GIS. This is not visible on your screen, but you have also set the WorldClim database as your default. You will use the climate data in Module 3.

3. Importing data into MaxEnt

After installing and starting MaxEnt, you will be asked to load a file with occurrence data (͞samples͟) and another file containing environmental variables (͞environmental layers͟).The occurrence data should be a .CSV file and the environmental layers in ASCII raster grids (each environmental variable represents one layer).
For example, the19 bioclimatic values representing the WorldClim datasets would appear as separate layers. After specifying anoutput file, you are ready to proceed to the next step of running the MaxEnt model. More on this topic in Module 3.

A full description of the procedure is available in Scheldeman and van Zonneveld (2010: 28–33).

Here is a quiz that will help you test your newly acquired knowledge. Once you have covered the content sections and completed the assigned readings, please answer the Data preparation and software selection quiz.

Continue to quiz  [33]

The next module in our research process is Climate change analysis  [34]. Let us begin!