Climate change and forests | Vegetation modelling

By: Rajiv K Chaturvedi, Ranjith Gopalakrishnan, Mathangi Jayaraman, Govindasamy Bala, N V Joshi, Raman Sukumar and N H Ravindranath
An assessment of the impact of climate change on forest ecosystems in India has been attempted in this paper, based on climate projections of the Regional Climate Model of the Hadley Centre and the dynamic global vegetation model IBIS for A2 and B2 scenarios. A forest vulnerability index for India has also been worked out based on the dynamic global vegetation modelling and observed datasets of forest density, forest biodiversity as well as model predicted vegetation type shift estimates for forested grids.
Forests

Climate is one of the most important determinants of vegetation patterns globally and has significant influence on the distribution, structure and ecology of forests. Several climate-vegetation studies have shown that certain climatic regimes are associated with particular plant communities or functional types. It is therefore logical to assume that changes in climate would alter the distribution of forest ecosystems. Based on a range of vegetation modelling studies, UN’s Intergovernmental Panel on Climate Change (IPCC) 2007 suggest potential forest dieback towards the end of this century and beyond, especially in tropics, boreal and mountain areas.

https://www.geographyandyou.com/climate-change/environment/iucn-study-finds-entire-ecosystems-affected-climate-change/

Assessments of potential climate change impacts on forests in India were based on the BIOME model (versions 3 and 4), which being an equilibrium model, does not capture the transient responses of vegetation to climate change. The recent study (Ravindranath et al. 2006) concludes that 77 per cent and 68 per cent of the forested grids in India are likely to experience shift in forest types for climate change under A2 and B2 scenarios, respectively (refer to page 13 for special report on emission scenarios – SRES). In addition, there have been two regional studies, the first focusing on potential climate change impacts on forests in Himachal Pradesh (Deshingkar 1997) and a second in the Western Ghats (Ravindranath et al. 1997). These studies indicated moderate to large scale shifts in vegetation types with implications for forest dieback and biodiversity. The studies conducted for India so far have faced several limitations, e.g., coarse resolution of input data; and, use of BIOME, an equilibrium model with limited capability in categorising plant functional types and dynamic representation of growth constraints.

Monsoon prediction

Impacts of climate change on forests have severe implications for people who depend on forest resources for their livelihoods. India is a mega biodiversity country where forests account for more than one fifth of the geographical area. With nearly 173,000 villages classified as forest villages, there is a large dependence of communities on forest resources. India has a huge afforestation programme of over 1.32 mha per annum, and more area is likely to be afforested under programmes ‘Green India Mission’ and ‘Compensatory Afforestation Fund Management and Planning Authority’ (CAMPA). It is thus imperative to assess the likely impacts of projected climate change on existing forests and afforested areas, and develop and implement adaptation strategies to enhance the resilience of forests to climate change.

Fig 1. Forest type, distribution and extent simulated by IBIS for the baseline case and A2 and B2 scenarios White areas represent non-forested grids. (VT - refers to Vegetation Types). The numbers refer to the following vegetation types 1: tropical evergreen forest / woodland, 2: tropical deciduous forest / woodland, 3. temperate evergreen broadleaf forest / woodland, 4: temperate evergreen conifer forest / woodland, 5: temperate deciduous forest / woodland, 6: boreal evergreen forest / woodland, 7: boreal deciduous forest / woodland, 8: mixed forest / woodland, 9: savanna, 10: grassland/ steppe, 11: dense shrubland, 12: open shrubland, 13: tundra, 14: desert, 15. polar desert / rock / ice
Fig 1. Forest type, distribution and extent simulated by IBIS for the baseline case and A2 and B2 scenarios
White areas represent non-forested grids. (VT – refers to Vegetation Types). The numbers refer to the following vegetation types 1: tropical evergreen forest / woodland, 2: tropical deciduous forest / woodland, 3. temperate evergreen broadleaf forest / woodland, 4: temperate evergreen conifer forest / woodland, 5: temperate deciduous forest / woodland, 6: boreal evergreen forest / woodland, 7: boreal deciduous forest / woodland, 8: mixed forest / woodland, 9: savanna, 10: grassland/ steppe, 11: dense shrubland, 12: open shrubland, 13: tundra, 14: desert, 15. polar desert / rock / ice

Status of forests in India

According to the Forest Survey of India (FSI) ‘all lands, more than one hectare in area, with a tree canopy density of more than 10 per cent is defined as forest’ (FSI 2009). The status of forests and forest management systems contribute to the vulnerability of forests to climate change. The Forest Survey of India has been periodically estimating the forest cover in India since 1987 using remote sensing techniques. In addition to forest cover, FSI has also included the tree cover in its 2001, 2003, 2005, and 2007 assessments.

Indian forests are extremely diverse and heterogeneous. Classification of Indian forest types is available from two main sources – Forest Survey of India (FSI 2001) and Champion and Seth (1968).  Due to forest heterogeneity, Forest Survey of India’s classification scheme has a pan-Indian ‘miscellaneous forest’ category (with no dominant species), which accounts for 63 per cent of forest area. This large miscellaneous category makes the FSI classification rather unattractive for further analysis. However, Champion and Seth (1968) classify Indian forests into 16 distinct forest types, prompting us to opt for the Champion and Seth classification for further analysis. The distribution of forest types in India according to Champion and Seth (1968) is shown in Fig 1.

Methods

The impacts of climate change on forests in India are assessed based on the changes in area under different forest types, shifts in boundary of forest types and net primary productivity (NPP). Data sets selected were: (i) spatial distribution of current climatic variables, (ii) similar data for future climate projected by relatively high resolution regional climate models for two different climate change scenarios, and (iii) vegetation types, NPP and carbon stocks as simulated by the dynamic model IBIS V.2 (Integrated Biosphere Simulator).

Article 10 Fig 2

Vegetation modelling

The dynamic vegetation model IBIS is designed around a hierarchical, modular structure. The model is broken into four modules – land surface module, vegetation phenology module, carbon balance module and vegetation dynamics module. These modules, though operating at different time steps, are integrated into a single physically consistent model that may be directly incorporated within AGCMs (atmospheric general circulation models). For example, IBIS is currently incorporated into two AGCMs namely GENESIS-IBIS and CCM3-IBIS. The model allows an understanding of different light and water regimes – enhancing comprehension of competition for sunlight and soil moisture which determines the geographic distribution of plant functional types and the relative dominance of trees and grasses, evergreen and deciduous phenologies, broadleaf and conifer leaf forms, and C3 and C4 photosynthetic pathways.

Input data

IBIS requires a range of input parameters, primarily climatological and soil characteristics. The main climatological parameters used are: monthly mean cloudiness (per cent), monthly mean precipitation rate (mm/day), monthly mean relative humidity (per cent), monthly minimum, maximum and mean temperature (C) and wind speed (m/s); while the main soil factor used is texture (i.e. percentage of sand, silt and clay). The model also requires topographical information.

Observed climatology is obtained from Climatic Research Unit (CRU), while soil data was obtained from International Geosphere-Biosphere Programme (IGBP). For climate change projections, RCM outputs from Hadley centre model HadRM3 were used.  The climate variables for future scenarios were obtained using the method of anomalies. Briefly, this involved computing the difference between projected values for a scenario and the control run of the HadRM3 model, and adding this difference to the value corresponding to the current climate as obtained from CRU climatology.  Climate data operators (CDO) software was used for the data editing and climate data analysis tool (CDAT) for data processing and generation of various maps and plots.

Fig 3. Percentage of forest grids undergoing vegetation change by 2085 under A2 and B2 scenarios according to forest types
Fig 3. Percentage of forest grids undergoing vegetation change by 2085 under A2 and B2 scenarios according to forest types

Selection of forested grids

Digitised forest map of India (FSI 2001) was used to determine the spatial location of all forested areas. This map was based on high resolution mapping (2.5 by 2.5 inch), wherein India was divided into over 165,000 grids. Out of these, 35,899 grids were marked as forested grids (along with forest density and forest type). Further, the forest grids were classified into three categories as per forest density: ‘very dense forests’ with crown density above 70 per cent; ‘moderately dense forest’ with crown density between 40 and 70 per cent; and, ‘open forest’ with crown density between 10 and 40 per cent.

Scenarios of climate change

SRES scenario A2 (740 ppm by 2085) is selected as one of the scenarios. However, since a more constrained emission pathway may emerge as a result of global mitigation actions, we also chose B2 scenario (575 ppm by 2085) in this study. The results were then compared with the ‘baseline’ (also referred to as reference or control case) scenario, which represents the simulation using the 1961-91 observed climatology.

Impacts of climate change on forest types and vegetation modelling

Changes in the distribution of forests

The vegetation distribution simulated by IBIS for baseline, A2 and B2 scenario in the forested grids are shown in Fig 1. It is noticed that there is an expansion of tropical evergreen forests (IBIS vegetation type 1) in eastern India plateau for both A2 and B2 scenarios. The same trend can be observed in the Western Ghats. It is interesting to note that there is almost no vegetation type change in the  northeast.  Further, there is a slight expansion of forests into the western part of central India. Overall, there is negligible difference between forest extents predicted for the future in A2 and B2 scenarios except that forest expansion is higher in the western part of central India in the A2 scenario. This could be attributed to higher precipitation levels in A2 scenario relative to B2 in this region. One caveat to the expansion trend of forests (like tropical evergreen) is the assumption that forests are not fragmented, and there is no dearth of seed dispersing agents. In the real world, forests are indeed fragmented, and, seed dispersal may not be efficient in the view of loss or reduction in number of dispersal agents due to human habitation pressures and climate change. As the population of seed dispersing agents decline, predicted forest expansion according to the vegetation modelling is not guaranteed.

Vulnerability index for India’s forests

Forests in India are already subjected to multiple stresses including over extraction, insect outbreaks, livestock grazing, forest fires and other anthropogenic pressures. Climate change is an additional one. Disturbed and fragmented, forests and monoculture forests are likely to be more vulnerable, to climate change. Therefore, a vulnerability index, Fig 2, has been developed to assess the risk factor of different forest types and regions. The various vulnerability index classes were defined by spatially combining information on forest diversity (monoculture versus natural forest), forest density (an indicator of degradation) and IBIS vegetation type change estimates for the forest grids under A2 scenario. For example, if a particular forest grid had monoculture vegetation, a low forest density (or higher levels of degradation) and if there was a vegetation type shift in the future as predicted by IBIS, then this grid point is given the highest vulnerability index of 7. The analysis thus achieved, points towards nearly 39 per cent of forested grids being vulnerable to climate change in India. The forests in central India are highly vulnerable. There are pockets of vulnerable forests surrounded by non vulnerable regions in that area.

A significant part of the Himalayan biodiversity hotspot that stretches along the north western part of India along the states of Punjab, Jammu and Kashmir and Himanchal Pradesh is projected to be highly vulnerable, mostly attributable to the higher elevation of these regions.  Our studies on vegetation modelling have shown that these regions will experience increased levels of warming.

Northern and central parts of the Western Ghats also seem to be significantly vulnerable to climate change. Northern parts of the Western Ghats contain significant extent of open forests, which drive up the vulnerability score. High values of the index in the central part of the Ghats are likely caused by the negligible precipitation increase over there (with more than 3°C rise in temperature). Forests in the southern part of the Western Ghats appear to be quite resilient as forests in this region are less fragmented, more diverse and they also support tropical wet evergreen forests which, according to IBIS simulations for vegetation modelling, are likely to remain stable. In the northeast of India, there are relatively few areas that have a high vulnerability index. This low vulnerability index in this regions is because climate is predicted to get hotter and wetter, which is conducive to the existing vegetation types – tropical evergreen forests.

The article is an extract of a previously online published work in Springer Science and Business Media BV in August 2010.

Leave a Reply

Your email address will not be published. Required fields are marked *