Global climate change is arguably the most dominant ecological problem of this generation. Extreme environmental events, shifting historic climate patterns and increased pressures on environmental services due to socio-ecological changes (i.e., deforestation and urban sprawl with population growth) continue to threaten the sustainability of economies, food supplies, and stocks of ecological resources.

As a result, the quantification of future changes in local-scale climate and agricultural production is becoming increasingly important to policy makers and land managers when attempting to develop mitigation planning strategies. CAVEAT is designed to satisfy such requirements and provide a spatially descriptive tool for decision-making, public policy building and the development of adaptation options.

Researchers at Trent University in Peterborough Ontario Canada have recently released Climate Change and Agricultural Visualization and Estimation Analysis Tool (CAVEAT) version 1.3. The tool has been developed as a proof of concept customizable decision support system for the analysis, display, and mapping of the effects of climate change at national and local scales.

CAVEAT is essentially the primary interface provided for the dissemination of information resulting from an intricate and complex modelling process, including: climate data produced by coupled atmosphere-ocean general circulation models (AOGCMs) adopted within the Intergovernmental Panel on Climate Change (IPCC) Fourth Assessment Report (AR4) and forced by paths of global development described by the IPCC – Special Report on Emission Scenarios (IPCC-SRES), crop production as simulated by the Food and Agriculture Organization of the United Nations (FAO) AquaCrop growth model, and a simulation calibration process complemented by a three phased recursive downscaling procedure for local scale spatially-continuous predictions.

The tools backend processing methodology was designed to be easily portable for any region of the world, at flexible time horizons, for any crop type, and for any scenario of global development. Currently the tool is customized to support predicted maize (Zea mays L.) production in Mexico for the year 2050 assuming the A1B scenario of global development, and the GFDL “CM2.1” AOGCM of global climate. Below in Figure 1 is a highly generalized illustration of the methodology undertaken to predict and communicate maize production.

Figure 1. An overview: Approach to maize (Zea mays L.) yield prediction in 2050 at the local-scale.

Further methodological inquires can be addressed by contacting the Trent University research team directly.


Metadata - *Beta 1.3*

CAVEAT at present is a prototype that provides basic functionality and proof of concept. Currently, the tool supports climate and agricultural predictions made assuming one path of global development (“A1B”), integrated within one AOGCM (the Geophysical Fluid Dynamics Laboratory - GFDL “CM 2.1” model), at one time horizon (the year 2050).

Future projected variables which are currently supported by CAVEAT include:

CAVEAT also supports the visualization of historic monthly climate station data (specifically, temperature and precipitation) acquired from the University of East Anglia’s Climate Research Unit TS 2.1 dataset averaged over the baseline period (1961-1990).

Calculated change or difference maps are developed from this baseline to provide renditions which serve to isolate regions of severe mean precipitation or temperature change projected to occur by the year 2050. Baseline and change maize yield maps are not yet supported.

All climatic maps are produced at a 0.5 degree resolution and maize predictions made for the year 2050 are produced at a 2,500 meter resolution.

All data currently supported by CAVEAT is available in KMZ format from the Export page.


Forcing Scenario

The selection of a forcing scenario (or SRES-scenario) within CAVEAT essentially identifies which path of development the Earth will follow into the future. This effectively allows for the customized display of the anticipated impacts of climate change under variable developmental trajectories. These scenarios are used to parameterize (or force) climate models for the simulation of atmospheric circulation patterns and spatial prediction of major climatic variables.

Currently (beta v1.3) only the “A1B” scenario is supported.


Since 1992 the Intergovernmental Panel on Climate Change (IPCC) has been standardizing, publishing, and subsequently enhancing a collection of socio-economic emissions scenarios with the specific aim to foster climate change research. Each update is directly related to a better understanding of the latest economic restructuring, the current rates and trends of technological change, and improved/extended emission baselines; most recently concluding in 2000 with the publishing of the Special Report on Emission Scenarios – SRES (to be replaced in 2014 with Representative Concentration Pathways). As the scenarios have advanced they have allowed progressive insight into the relationships which dictate environmental quality, specifically those related to social, technological, and economic choices on a global stage and have lead to the development of four major IPCC impact assessment reports (available at:

SRES-scenarios can be essentially thought of as plausible perspectives of how greenhouse gas and aerosol concentrations might unfold into the future, based on the progression of global demographics, socio-economic development, and technological change. The production of these scenarios is achieved via comprehensive evaluations of emission driving forces in scenario literature, variable avenues of modelling global development, and an “open process” which sought out independent participatory research and feedback. However, there currently exists no way of knowing with certainty future global development; therefore, a series of ‘storylines’ were developed as equally valid with no assigned probability of occurrence (Nakicenovic et al., 2000).

The four major storylines of possible global development include: A1, A2, B1, and B2, graphically shown in Figure 2. As the scenarios advance in a temporal dimension each describes a different demographic, social, economic, technological, and environmental future which progressively and irrevocably deviates from one another. The four storylines are predicated on two divergent tendencies; one contrasting strong economic vs. strong environmental values, and two, contrasting increasing globalization vs. increasing regionalization (Carter et al., 2007).

Figure 2. SRES-scenario tree illustrating the major socio-economic divergent scenario tendencies (Carter et al., 2007).

The following descriptions are summaries of those described in the SRES by Nakicenovic et al. (2000):

The A1 storyline describes a world of accelerated economic growth, a population expansion which climaxes mid-century, rapid infusion of exceedingly efficient technologies, increased infrastructural developments, reduced cultural divisions and increased social communication and acceptance, and significant declines in the variation of regional per capita income. The A1 storyline is unique in that it further breaks down into three scenario groups which illustrate distinct variations of technological evolution specifically related to energy production: 1.- A1FI, fossil fuel intensive energy, 2.- A1T, non-fossil fuel based energy, and 3.- A1B, a balance between fossil and non-fossil fuel sources where similar usage and technological evolution rates dominate.

The A2 storyline and scenario group describes a heterogeneous world of isolation and cultivation of local identities and governments. Here regional fertility rates homogenize slowly resulting in rising populations. Since the underlying theme is essentially regional independence, economies maintain autonomy resulting in the depression of per-capita income growth rates and technological evolution relative to other storylines.

The B1 storyline and scenario group describes a homogenous world with a mid-century global population climax similar to A1. However, here there is more rapid progress towards service and information based economies and increased development of clean and resource-efficient technologies. The underlying theme of the B1 storyline is global responses to economic, social, and environmental sustainability.

The B2 storyline and scenario group describes a heterogeneous world with regional responses to economic, social, and environmental sustainability. Global populations rise at a rate lower than A2, and rates of technological change are depressed compared to A1 and B1 with similar trends in economic development. The underlying theme here is local and regional solutions to environmental conservation and social equality.

Of these scenarios the “A1B” SRES-scenario has been integrated and is available for selection within CAVEAT (v1.3). This scenario was selected as it has been commonly adopted as a "business-as-usual" (BAU) global developmental pathway in the scenario litterature and since it resembles and maintains developmental trajectories similar to current trends (Stroeve et al., 2007; Robock et al., 2008; Donner, 2009; Caesar, 2010; Pardaens et al., 2011). That being said, the scenario has no more probability of occurring than any of the other scenarios (Nakicenovic et al., 2000), but may provide a best estimate of future greenhouse gas emissions.


Climate Model

The selection of a climate model within CAVEAT essentially allows the user the ability to view the impacts of climate change under wet, medium, or dry conditions at the time horizon and region of study. Depending on agricultural requirements, this selection can translate into best or worst case impact predictions (e.g. yield production) as the climate output by these models is subsequently used to parameterize a crop growth simulation model (AquaCrop).

Currently (beta v1.3) only the Geophysical Fluid Dynamics Laboratory (GFDL) “CM 2.1” (“wet”) atmosphere-ocean general circulation climate model is supported.


Global climate is a very large and complicated system; there are always uncertainties when attempting to predict future conditions. One of the primary instruments available to researches for simulating global climate and predicting the effects of anthropogenic modifications to the atmosphere are general circulation models or GCMs. These models comprehensively describe the climate system in three dimensions by representing its dynamics, physical processes, interactions, and feedbacks (Ruosteenoja et al., 2003). At their core, all GCMs are in essence numerical highly complex mathematical models programmed to run on the world’s most advanced computers which simulate the Earth’s atmospheric circulation processes through fundamental concepts of conservation of energy, momentum, motion, and adherence to the law of ideal gases (McGuffie and Henderson-Sellers, 2005). GCMs can be conceptualized via three distinct components, namely: 1.- algorithm dynamics representing horizontal and vertical movements of mass and energy within and between columns of atmosphere and ocean; 2.- physics routines encapsulating comparatively small scale convective and radiative fluxes; and 3.- surface processes including energy and mass interactions of the atmosphere, sea, ice, and land (McGuffie and Henderson-Sellers, 2005). These interactions are parameterized within GCMs to model climate by first segmenting the globe in three dimensions, typically into horizontally equidistant grid cells between 250 and 600km in size, 10 to 20 vertical layers in the atmosphere, and up to 30 vertical layers in the oceans. The GCM then simulates all interactions and movements of mass and energy within and between grid cells, as graphically depicted in Figure 3. Cell size, i.e. model spatial resolution, is directly related to the input data, model research objectives, and available computing power (Carter et al., 2007).

Figure 3. Conceptual structure of a coupled atmosphere-ocean general circulation model (Viner and Hulme, 1997).

As GCMs have evolved they have become progressively more comprehensive and dynamic, incorporating the principles of regional weather and hydrological models (McGuffie and Henderson-Sellers, 2005). Moreover, significant advancements were also made via the incorporation of vegetation canopies and snow surface models (Lynch-Stieglitz., 1994; Verseghy, 1996), the inclusion of clouds and their effects on solar and longwave radiation (IPCC, 1996), and of oceanic circulation models and their important role on heat transfers and CO2 fluxes (Gordon et al., 2000). Hence, most GCMs incorporate at least three distinct component sub-models, namely, atmosphere, land-surface, and ocean. GCMs which dynamically link the ocean to atmosphere are the most advanced and also commonly referred to as coupled atmosphere-ocean general circulation models (AOGCMs). These models are able to integrate ocean currents and the corresponding global transfers of greenhouse gases, heat, momentum, and moisture from the atmosphere thus allowing for the simulation of time lags (or feedbacks) between atmospheric composition change and the corresponding change in climate (Manabe et al., 1991; Carter et al., 2007). These AOGCMs also often incorporate a fourth cryospheric sub-model component, parameterizing changes in global albedo and fresh melt water additions to the system allowing for the most realistic (to date) long term climate change simulation responses to increased greenhouse gases (Carter et al., 2007).

When these AOGCMs are forced by SRES-scenarios of future greenhouse gas emissions they are able to simulate future climate and develop plausible scenarios of regional climate change. Such models were utilized in the IPCC Fourth Assessment Report - AR4 (IPCC, 2007) and are the only models given focus within CAVEAT.

The climate simulated by these AOGCMs were averaged for the time horizon and Mexican region of study to identify those climate models which had simulation tendencies toward wet, medium, or dry conditions. The “wet” Geophysical Fluid Dynamics Laboratory (GFDL) “CM 2.1” model was selected and used to simulate climate for integration within agricultural yield predictions produced by AquaCrop.


Time Horizon

The selection of a time horizon within CAVEAT allows the user the ability to view the impacts of climate change at a specified future time period of study. This functionality was specifically designed to assist mitigation planning providing impact analyses in the short (2020), medium (2050) and long-term future (2100).

Currently (beta v1.3) only the medium time horizon is supported, projecting the impacts of climate change into the year 2050.


The backend methodology which has been developed by researchers at Trent University to make impact predictions for the year 2050 can be easily adjusted for any other time horizon.


Display Variable

The selection of a display variable within CAVEAT allows the user the ability to customize their view of the anticipated impacts of climate change, from atmospheric characteristics to predicted agricultural production. This effectivly provides users with a context for interpreting the impacts of climate change. By mapping and providing projected atmospheric climate patterns and their changes from observed distributions, a major source of changes in agricultural production can be spatially understood and thereby provide a frame of reference for mitigation planning.

Currently (beta v1.3) mean temperature and precipitation atmospheric variables are provided to complement predicted maize yields.


Spatially-explicit maize (Zea mays L.) yield predictions (in 2050) were developed as described in Figure 1 (please contact the research team at Trent University for further information). Temperature and precipitation maps were produced via three different processes depending on their specification.

All “baseline” observed climate data were downloaded from the Consultative Group for International Agriculture Research - Consortium for Spatial Information (CGIAR-CSI) GeoPortal - CRU TS 2.1 Global Climate Database ( The CRU TS 2.1 is a terrestrial climate dataset based on historic climate station measurements produced by the Climate Research Unit (CRU) of the University of East Anglia on a global equidistant grid, extending for the period 1901 – 2002. The displayed baseline data within CAVEAT averages thirty years (1961-1990) of monthly values provided by this dataset and within the Mexican study site to provide a basis from which the impacts of climate change can be referenced.

All "future" climate data displayed within CAVEAT were produced by the AOGCMs of the AR4 and correspond to the user’s forcing scenario, climate model and time horizon selection. This data was downscaled via minimum curvature bi-cubic spline to match that of the CRU baseline for comparisons and change analyses. A bi-cubic spline spatial-interpolation was chosen to downscale since the baseline gridded data had also been downscaled via smoothing algorithm (Giorgi et al., 2004), and since temperature and precipitation monthly average climate trends at the baseline coarse resolution (0.5o) are likely to transition gradually rather than be highly subjective to local influences. Note: the prediction of maize yields did not adopt this method of downscaling, contact the developers for further information.

All climate data displayed at the “change” specification within CAVEAT therefore represents the difference between the climate generated at the baseline and future time periods.

Future versions of CAVEAT will include many other variables to provide users with necessary information to conceptually frame and establish causality with the simulated agricultural impacts, and thus allow for effective policy development (e.g., PROMAF - Gonzalez et al., 2010), infrastructural planning (e.g., irrigation and reservoir development, etc.), and further impact assessment studies (e.g., river discharge rates and erosion hotspots, etc.).



The specification selection option within CAVEAT tells the tool whether to output future, baseline, or change mapping products of the respective display variable selected. If the baseline selection is made, logically the selection of a forcing scenario, climate model, and time horizon is not necessary.

Of the options currently (beta v1.3) available only the baseline and change maize (Zea mays L.) yield mapping products are not yet supported. If these are selected an error prompt will appear.


For further inquisitions, methodological descriptions and full references please contact the research development team at Trent University directly.