Normalized Difference Vegetation Index (normalized + difference_vegetation_index)

Distribution by Scientific Domains
Distribution within Life Sciences

Selected Abstracts

Are local weather, NDVI and NAO consistent determinants of red deer weight across three contrasting European countries?

Abstract There are multiple paths via which environmental variation can impact herbivore ecology and this makes the identification of drivers challenging. Researchers have used diverse approaches to describe the association between environmental variation and ecology, including local weather, large-scale patterns of climate, and satellite imagery reflecting plant productivity and phenology. However, it is unclear to what extent it is possible to find a single measure that captures climatic effects over broad spatial scales. There may, in fact, be no a priori reason to expect populations of the same species living in different areas to respond in the same way to climate as their population may experience limiting factors at different times of the year, and the forms of regulation may differ among populations. Here, we examine whether the same environmental indices [seasonal Real Bioclimatic Index (RBI), seasonal Normalized Difference Vegetation Index (NDVI) and winter North Atlantic Oscillation (NAO)] influence body size in different populations of a large ungulate living in Mediterranean Spain, Western Scotland and Norway. We found substantial differences in the pattern of weight change over time in adult female red deer among study areas as well as different environmental drivers associated with variation in weight. The lack of general patterns for a given species at a continental scale suggest that detailed knowledge regarding the way climate affects local populations is often necessary to successfully predict climate impact. We caution against extrapolation of results from localized climate,population studies to broad spatial scales. [source]

Land-use impact on ecosystem functioning in eastern Colorado, USA

J. M. Paruelo
Abstract Land-cover change associated with agriculture has had an enormous effect on the structure and functioning of temperate ecosystems. However, the empirical evidence for the impact of land use on ecosystem functioning at the regional scale is scarce. Most of our knowledge on land-use impact has been derived from simulation studies or from small plot experiments. In this article we studied the effects of land use on (i) the seasonal dynamics and (ii) the interannual variability of the Normalized Difference Vegetation Index (NDVI), a variable linearly related to the fraction of the photosynthetically active radiation (PAR) intercepted by the canopy. We also analysed the relative importance of environmental factors and land use on the spatial patterns of NDVI. We compared three cultivated land-cover types against native grasslands. The seasonal dynamics of NDVI was used as a descriptor of ecosystem functioning. In order to reduce the dimensionality of our data we analysed the annual integral (NDVI-I), the date of maximum NDVI (DMAX) and the quarterly average NDVI. These attributes were studied for 7 years and for 346 sites distributed across eastern Colorado (USA). Land use did modify ecosystem functioning at the regional level in eastern Colorado. The seasonal dynamics of NDVI, a surrogate for the fraction of PAR intercepted by the canopy, were significantly altered by agricultural practices. Land use modified both the NDVI integral and the seasonal dynamics of this spectral index. Despite the variability within land-cover categories, land use was the most important factor in explaining regional differences of the NDVI attributes analysed. Within the range of environmental conditions found in eastern Colorado, land use was more important than mean annual precipitation, mean annual temperature and soil texture in determining the seasonal dynamics of NDVI. [source]

Loop migration in adult marsh harriers Circus aeruginosus, as revealed by satellite telemetry

Raymond H. G. Klaassen
Loop migration among birds is characterized by the spring route lying consistently west or east of the autumn route. The existence of loops has been explained by general wind conditions or seasonal differences in habitat distribution. Loop migration has predominantly been studied at the population level, for example by analysing ring recoveries. Here we study loop migration of individual marsh harriers Circus aeruginosus tracked by satellite telemetry. We show that despite a generally narrow migration corridor the harriers travelled in a distinct clockwise loop through Africa and southern Europe, following more westerly routes in spring than in autumn. We used the Normalized Difference Vegetation Index (NDVI) to identify potential feeding habitat in Africa. Suitable habitat seemed always more abundant along the western route, both in spring and autumn, and no important stopover site was found along the eastern route. Observed routes did thus not coincide with seasonal variation in habitat availability. However, favourable habitat might be more important during spring migration, when the crossing of the Sahara seems more challenging, and thus habitat availability might play an indirect role in the harriers' route choice. Grid-based wind data were used to reconstruct general wind patterns, and in qualitative agreement with the observed loop marsh harriers predominantly encountered westerly winds in Europe and easterly winds in Africa, both in autumn and in spring. By correlating tail- and crosswinds with forward and perpendicular movement rates, respectively, we show that marsh harriers are partially drifted by wind. Thus, we tentatively conclude that wind rather than habitat seems to have an overriding effect on the shape of the migration routes of marsh harriers. General wind conditions seem to play an important role also in the evolution of narrow migratory loops as demonstrated for individual marsh harriers. [source]

Expansion of human settlement in Kenya's Maasai Mara: what future for pastoralism and wildlife?

Richard H. Lamprey
Abstract Aim, Wildlife and pastoral peoples have lived side-by-side in the Mara ecosystem of south-western Kenya for at least 2000 years. Recent changes in human population and landuse are jeopardizing this co-existence. The aim of the study is to determine the viability of pastoralism and wildlife conservation in Maasai ranches around the Maasai Mara National Reserve (MMNR). Location, A study area of 2250 km2 was selected in the northern part of the Serengeti-Mara ecosystem, encompassing group ranches adjoining the MMNR. Emphasis is placed on Koyake Group Ranch, a rangeland area owned by Maasai pastoralists, and one of Kenya's major wildlife tourism areas. Methods, Maasai settlement patterns, vegetation, livestock numbers and wildlife numbers were analysed over a 50-year period. Settlement distributions and vegetation changes were determined from aerial photography and aerial surveys of 1950, 1961, 1967, 1974, 1983 and 1999. Livestock and wildlife numbers were determined from re-analysis of systematic reconnaissance flights conducted by the Kenya Government from 1977 to 2000, and from ground counts in 2002. Corroborating data on livestock numbers were obtained from aerial photography of Maasai settlements in 2001. Trends in livestock were related to rainfall, and to vegetation production as indicated by the seasonal Normalized Difference Vegetation Index. With these data sets, per capita livestock holdings were determined for the period 1980,2000, a period of fluctuating rainfall and primary production. Results, For the first half of the twentieth century, the Mara was infested with tsetse-flies, and the Maasai were confined to the Lemek Valley area to the north of the MMNR. During the early 1960s, active tsetse-control measures by both government and the Maasai led to the destruction of woodlands across the Mara and the retreat of tsetse flies. The Maasai were then able to expand their settlement area south towards MMNR. Meanwhile, wildebeest (Connochaetes taurinus) from the increasing Serengeti population began to spill into the Mara rangelands each dry season, leading to direct competition between livestock and wildlife. Group ranches were established in the area in 1970 to formalize land tenure for the Maasai. By the late 1980s, with rapid population growth, new settlement areas had been established at Talek and other parts adjacent to the MMNR. Over the period 1983,99, the number of Maasai bomas in Koyake has increased at 6.4% per annum (pa), and the human population at 4.4% pa. Over the same period, cattle numbers on Koyake varied from 20,000 to 45,000 (average 25,000), in relation to total rainfall received over the previous 2 years. The rangelands of the Mara cannot support a greater cattle population under current pastoral practices. Conclusions, With the rapid increase in human settlement in the Mara, and with imminent land privatization, it is probable that wildlife populations on Koyake will decline significantly in the next 3,5 years. Per capita livestock holdings on the ranch have now fallen to three livestock units/reference adult, well below minimum pastoral subsistence requirements. During the 1980s and 90s the Maasai diversified their livelihoods to generate revenues from tourism, small-scale agriculture and land-leases for mechanized cultivation. However, there is a massive imbalance in tourism incomes in favour of a small elite. In 1999 the membership of Koyake voted to subdivide the ranch into individual holdings. In 2003 the subdivision survey allocated plots of 60 ha average size to 1020 ranch members. This land privatization may result in increased cultivation and fencing, the exclusion of wildlife, and the decline of tourism as a revenue generator. This unique pastoral/wildlife system will shortly be lost unless land holdings can be managed to maintain the free movement of livestock and wildlife. [source]

Remote sensing of protected areas to derive baseline vegetation functioning characteristics

Martín F. Garbulsky
Abstract: Question: How can we derive baseline/reference situations to evaluate the impact of global change on terrestrial ecosystem functioning? Location: Main biomes (steppes to rain forests) of Argentina. Methods: We used AVHRR/NOAA satellite data to characterize vegetation functioning. We used the seasonal dynamics of the Normalized Difference Vegetation Index (NDVI), a linear estimator of the fraction of the photosynthetic active radiation intercepted by vegetation (fPAR), and the surface temperature (Ts), for the period 1981,1993. We extracted the following indices: NDVI integral (NDVI -I), NDVI relative range (Rrel), NDVI maximum value (Vmax), date of maximum NDVI (Dmax) and actual evapotranspiration. Results: fPAR varied from 2 to 80%, in relation to changes in net primary production (NPP) from 83 to 1700 g.m- 2.yr -1. NDVI -I, Vmax and fPAR had positive, curvilinear relationships to mean annual precipitation (MAP), NPP was linearly related to MAP. Tropical and subtropical biomes had a significantly lower seasonality (Rrel) than temperate ones. Dmax was not correlated with the defined environmental gradients. Evapotranspiration ranged from 100 to 1100 mm.yr -1. Interannual variability of NDVI attributes varied across the temperature and precipitation gradients. Conclusions: Our results may be used to represent baseline conditions in evaluating the impact of land use changes across environmental gradients. The relationships between functional attributes and environmental variables provide a way to extrapolate ecological patterns from protected areas across modified habitats and to generate maps of ecosystem functioning. [source]

Influence of land-use types and climatic variables on seasonal patterns of NDVI in Mediterranean Iberian ecosystems

P. Durante
Abstract Question: What is the influence of management on the functioning of vegetation over time in Mediterranean ecosystems under different climate conditions? Location: Mediterranean shrublands and forests in SE Iberia (Andalusia). Methods: We evaluated the Normalized Difference Vegetation Index (NDVI) for the 1997-2002 time series to determine phenological vegetation patterns under different historical management regimes. Three altitudinal ranges were considered within each area to explore climate × management interactions. Each phenological pattern was analysed using time series statistics, together with precipitation (monthly and cumulative) and temperature. Results: NDVI time series were significantly different under different management regimes, particularly in highly transformed areas, which showed the lowest NDVI, weakest annual seasonality and a more immediate phenological response to precipitation. The NDVI relationship with precipitation was strongest in the summer-autumn period, when precipitation is the main plant growth-limiting factor. Conclusions: NDVI time series analyses elucidated complex influences of land use and climate on ecosystem functioning in these Mediterranean ecosystems. We demonstrated that NDVI time series analyses are a useful tool for monitoring programmes because of their sensitivity to changes, ease of use and applicability to large-scale studies. [source]

Linking vegetation heterogeneity and functional attributes of temperate grasslands through remote sensing

Roxana Aragón
Abstract Question: How are plant communities of the Flooding Pampa grasslands spatially distributed? How do canopy dynamics of the different communities vary among seasons and years? Location: Buenos Aires province, Argentina. Methods: We characterized the distribution of communities through a supervised classification based on four Landsat 5TM images. We sampled species composition of 200 sites, with 130 of them corresponding to natural communities. Of the sampling areas 60% were used to classify, and the remaining areas to assess classification accuracy. We characterized the seasonal and interannual variability of canopy dynamics using NDVI (Normalized Difference Vegetation Index) data provided by MODIS /Terra images. Results: Overall accuracy of the classification was satisfactory. The resulting maps showed a landscape formed by a matrix of extended lowlands with small patches of mesophytic and humid mesophytic meadows. The October scene (near the peak of productivity) was particularly important in discriminating among communities. The seasonal pattern of NDVI differed among communities and years. Mesophytic meadows had the highest NDVI mean and the lowest interannual coefficient of variation, halophytic steppes had the lowest mean, and vegetated ponds were the most variable. Conclusions: These grasslands have a fine-grained heterogeneity at the landscape scale. Each plant community has distinct seasonal and interannual canopy dynamics. These two features of grassland structure and functioning represent key information for rangeland management that may be obtained through a combination of minor field sampling and remote sensing. [source]

Temporal and spatial patterns of ecosystem functioning in protected arid areas in southeastern Spain

José M. Paruelo
Abstract We characterized the spatial variability and temporal dynamics of the photosynthetic active radiation absorbed (APAR) by the canopy, a descriptor of ecosystem functioning, in Cabo de Gata , Níjar Natural Park (CGNNP) (Spain). Ecosystem functioning was characterized for five landscape classes using the Normalized Difference Vegetation Index (NDVI) derived from NOAA/AVHRR LAC (1 km × 1 km) images. We also used a 19-year time series of NDVI PAL data (8 km × 8 km) to analyse the relationship APAR-precipitation inside and outside the park. The vegetation of CGNNP absorbed less than 20% of the incoming radiation. Plains intercepted 37% and hills 14% less photosynthetic active radiation than mountains, the most productive landscape of the park. CGNNP showed a well-defined growing season with a unique peak of APAR. Plains and piedmont, covered by annual vegetation displayed an earlier development of the leaf area index than the shrublands and grasslands typical of the other landscapes. APAR had a significant relationship with the sum of the precipitation of the current and two previous growing seasons, except for the plains. We found that the APAR of the areas more modified by humans (outside the park) showed a lower sensitivity to changes in precipitation than those under protection. The differences were higher if the accumulated precipitation of the previous three growing seasons was considered. The description of such differences in the response of absorbed PAR to water availability are proposed as the base of a monitoring system for semi-arid and arid areas. [source]

The influence of increasing population size and vegetation productivity on elephant distribution in the Kruger National Park

Abstract Decisions to reduce the impacts of large herbivores on biodiversity in protected areas are often based on controlling their numbers. However, numbers per se may not be the foremost consideration when managing impacts. This is because density-related changes in distribution can also affect habitat utilization and hence, impact. In this study we tested whether changes in the distribution of African elephants are associated with increasing population size. We used spatially explicit count data collected during the dry seasons from 1998 to 2004 in South Africa's Kruger National Park. We did this at five spatial scales and in landscapes defined by vegetation, geology, climate and soils. We then investigated whether observed distributions and grid-cell-specific densities were associated with the remotely sensed Normalized Difference Vegetation Index (NDVI) as a measure of productivity and therefore food resource availability at the landscape scale. Consistent with density mediated changes, we found that elephant grid-cell occupancy increased with population size, while grid-cell-specific density became less variable. In addition, the combined distribution of bull groups and breeding herds became less clumped with increasing population size. We further found that within landscapes elephants were present on grid-cells with higher NDVI values, but that the influence of NDVI during the dry season on densities among landscapes may be weak. These results suggest that NDVI was more indicative of structural habitat choices such as woody vegetation than food availability per se. Our study highlights the need to consider factors other than population size alone when formulating management decisions to reduce large herbivore impacts on biodiversity in protected areas. [source]

Environmental factors and population fluctuations of Akodon azarae (Muridae: Sigmodontinae) in central Argentina

Abstract The aim of this work was to explore the relationship between population density of Akodon azarae (Muridae: Sigmodontinae) and climatic and environmental variables, and determine which of them are associated to within and among-year changes in rodent abundance in agro-ecosystems from south Córdoba, Argentina. The study was carried out in a rural area of central Argentina, from 1983 to 2003. Density was estimated as a relative density index (RDI). Temperature, precipitation and humidity were obtained from records of the National University of Rio Cuarto. Normalized Difference Vegetation Index (NDVI) and Land Surface Temperature were recorded from National Oceanic and Atmospheric Administration (1983,1998) and Landsat (1998,2003) imagery data sets. We performed simple correlations, multiple regressions and distributed lag analysis. Direct association of climatic and environmental variables with RDI was in general, low. The amount of variability in seasonal changes in density explained by climatic and environmental variables altogether varied from 10% to 70%. Seasonal population fluctuations were influenced by NDVI and rainfall with one and two seasons of delay. Autumn maximum density of the species was also associated with vegetation and rainfall of previous seasons. There also seemed to be an indirect influence of rainfall through vegetation given that we found a positive correlation between them. Results were consistent with basic aspects of the ecology of the species, such as its strong preference for highly covered areas, which provide food and protection from predators, likely increasing its reproductive success. Therefore, in the rural area central Argentina, A. azarae showed seasonal fluctuations with delayed influence of rainfall and vegetation and indirect effects of rainfall. [source]

Patterns of spatial autocorrelation of assemblages of birds, floristics, physiognomy, and primary productivity in the central Great Basin, USA

Erica Fleishman
ABSTRACT We fitted spatial autocorrelation functions to distance-based data for assemblages of birds and for three attributes of birds' habitats at 140 locations, separated by up to 65 km, in the Great Basin (Nevada, USA). The three habitat characteristics were taxonomic composition of the vegetation, physical structure of the vegetation, and a measure of primary productivity, the normalized difference vegetation index, estimated from satellite imagery. We found that a spherical model was the best fit to data for avifaunal composition, vegetation composition, and primary productivity, but the distance at which spatial correlation effectively was zero differed substantially among data sets (c. 30 km for birds, 20 km for vegetation composition, and 60 km for primary productivity). A power-law function was the best fit to data for vegetation structure, indicating that the structure of vegetation differed by similar amounts irrespective of distance between locations (up to the maximum distance measured). Our results suggested that the spatial structure of bird assemblages is more similar to vegetation composition than to either vegetation structure or primary productivity, but is autocorrelated over larger distances. We believe that the greater mobility of birds compared with plants may be responsible for this difference. [source]

A Geostatistical Analysis of Soil, Vegetation, and Image Data Characterizing Land Surface Variation

Sarah E. Rodgers
The elucidation of spatial variation in the landscape can indicate potential wildlife habitats or breeding sites for vectors, such as ticks or mosquitoes, which cause a range of diseases. Information from remotely sensed data could aid the delineation of vegetation distribution on the ground in areas where local knowledge is limited. The data from digital images are often difficult to interpret because of pixel-to-pixel variation, that is, noise, and complex variation at more than one spatial scale. Landsat Thematic Mapper Plus (ETM+) and Satellite Pour l'Observation de La Terre (SPOT) image data were analyzed for an area close to Douna in Mali, West Africa. The variograms of the normalized difference vegetation index (NDVI) from both types of image data were nested. The parameters of the nested variogram function from the Landsat ETM+ data were used to design the sampling for a ground survey of soil and vegetation data. Variograms of the soil and vegetation data showed that their variation was anisotropic and their scales of variation were similar to those of NDVI from the SPOT data. The short- and long-range components of variation in the SPOT data were filtered out separately by factorial kriging. The map of the short-range component appears to represent the patterns of vegetation and associated shallow slopes and drainage channels of the tiger bush system. The map of the long-range component also appeared to relate to broader patterns in the tiger bush and to gentle undulations in the topography. The results suggest that the types of image data analyzed in this study could be used to identify areas with more moisture in semiarid regions that could support wildlife and also be potential vector breeding sites. [source]

Debating the greening vs. browning of the North American boreal forest: differences between satellite datasets

Abstract A number of remote sensing studies have evaluated the temporal trends of the normalized difference vegetation index (NDVI or vegetation greenness) in the North American boreal forest during the last two decades, often getting quite different results. To examine the effect that the use of different datasets might be having on the estimated trends, we compared the temporal trends of recently burned and unburned sites of boreal forest in central Canada calculated from two datasets: the Global Inventory, Monitoring, and Modeling Studies (GIMMS), which is the most commonly used 8 km dataset, and a new 1 km dataset developed by the Canadian Centre for Remote Sensing (CCRS). We compared the NDVI trends of both datasets along a fire severity gradient in order to evaluate the variance in regeneration rates. Temporal trends were calculated using the seasonal Mann,Kendall trend test, a rank-based, nonparametric test, which is robust against seasonality, nonnormality, heteroscedasticity, missing values, and serial dependence. The results showed contrasting NDVI trends between the CCRS and the GIMMS datasets. The CCRS dataset showed NDVI increases in all recently burned sites and in 50% of the unburned sites. Surprisingly, the GIMMS dataset did not capture the NDVI recovery in most burned sites and even showed NDVI declines in some burned sites one decade after fire. Between 50% and 75% of GIMMS pixels showed NDVI decreases in the unburned forest compared with <1% of CCRS pixels. Being the most broadly used dataset for monitoring ecosystem and carbon balance changes, the bias towards negative trends in the GIMMS dataset in the North American boreal forest has broad implications for the evaluation of vegetation and carbon dynamics in this region and globally. [source]

Evaluating the impacts of climate and elevated carbon dioxide on tropical rainforests of the western Amazon basin using ecosystem models and satellite data

Abstract Forest inventories from the intact rainforests of the Amazon indicate increasing rates of carbon gain over the past three decades. However, such estimates have been questioned because of the poor spatial representation of the sampling plots and the incomplete understanding of purported mechanisms behind the increases in biomass. Ecosystem models, when used in conjunction with satellite data, are useful in examining the carbon budgets in regions where the observations of carbon flows are sparse. The purpose of this study is to explain observed trends in normalized difference vegetation index (NDVI) using climate observations and ecosystem models of varying complexity in the western Amazon basin for the period of 1984,2002. We first investigated trends in NDVI and found a positive trend during the study period, but the positive trend in NDVI was observed only in the months from August to December. Then, trends in various climate parameters were calculated, and of the climate variables considered, only shortwave radiation was found to have a corresponding significant positive trend. To compare the impact of each climate component, as well as increasing carbon dioxide (CO2) concentrations, on evergreen forests in the Amazon, we ran three ecosystem models (CASA, Biome-BGC, and LPJ), and calculated monthly net primary production by changing a climate component selected from the available climate datasets. As expected, CO2 fertilization effects showed positive trends throughout the year and cannot explain the positive trend in NDVI, which was observed only for the months of August to December. Through these simulations, we demonstrated that the positive trend in shortwave radiation can explain the positive trend in NDVI observed for the period from August to December. We conclude that the positive trend in shortwave radiation is the most likely driver of the increasing trend in NDVI and the corresponding observed increases in forest biomass. [source]

Human modification of the landscape and surface climate in the next fifty years

R. S. Defries
Abstract Human modification of the landscape potentially affects exchanges of energy and water between the terrestrial biosphere and the atmosphere. This study develops a possible scenario for land cover in the year 2050 based on results from the IMAGE 2 (Integrated Model to Assess the Greenhouse Effect) model, which projects land-cover changes in response to demographic and economic activity. We use the land-cover scenario as a surface boundary condition in a biophysically-based land-surface model coupled to a general circulation model for a 15-years simulation with prescribed sea surface temperature and compare with a control run using current land cover. To assess the sensitivity of climate to anthropogenic land-cover change relative to the sensitivity to decadal-scale interannual variations in vegetation density, we also carry out two additional simulations using observed normalized difference vegetation index (NDVI) from relatively low (1982,83) and high (1989,90) years to describe the seasonal phenology of the vegetation. In the past several centuries, large-scale land-cover change occurred primarily in temperate latitudes through conversion of forests and grassland to highly productive cropland and pasture. Several studies in the literature indicate that past changes in surface climate resulting from this conversion had a cooling effect owing to changes in vegetation morphology (increased albedo). In contrast, this study indicates that future land-cover change, likely to occur predominantly in the tropics and subtropics, has a warming effect governed by physiological rather than morphological mechanisms. The physiological mechanism is to reduce carbon assimilation and consequently latent relative to sensible heat flux resulting in surface temperature increases up to 2 °C and drier hydrologic conditions in locations where land cover was altered in the experiment. In addition, in contrast to an observed decrease in diurnal temperature range (DTR) over land expected with greenhouse warming, results here suggest that future land-cover conversion in tropics could increase the DTR resulting from decreased evaporative cooling during the daytime. For grid cells with altered land cover, the sensitivity of surface temperature to future anthropogenic land-cover change is generally within the range induced by decadal-scale interannual variability in vegetation density in temperate latitudes but up to 1.5 °C warmer in the tropics. [source]

The greening and browning of Alaska based on 1982,2003 satellite data

GLOBAL ECOLOGY, Issue 4 2008
David Verbyla
Abstract Aim To examine the trends of 1982,2003 satellite-derived normalized difference vegetation index (NDVI) values at several spatial scales within tundra and boreal forest areas of Alaska. Location Arctic and subarctic Alaska. Methods Annual maximum NDVI data from the twice monthly Global Inventory Modelling and Mapping Studies (GIMMS) NDVI 1982,2003 data set with 64-km2 pixels were extracted from a spatial hierarchy including three large regions: ecoregion polygons within regions, ecozone polygons within boreal ecoregions and 100-km climate station buffers. The 1982,2003 trends of mean annual maximum NDVI values within each area, and within individual pixels, were computed using simple linear regression. The relationship between NDVI and temperature and precipitation was investigated within climate station buffers. Results, At the largest spatial scale of polar, boreal and maritime regions, the strongest trend was a negative trend in NDVI within the boreal region. At a finer scale of ecoregion polygons, there was a strong positive NDVI trend in cold arctic tundra areas, and a strong negative trend in interior boreal forest areas. Within boreal ecozone polygons, the weakest negative trends were from areas with a maritime climate or colder mountainous ecozones, while the strongest negative trends were from warmer basin ecozones. The trends from climate station buffers were similar to ecoregion trends, with no significant trends from Bering tundra buffers, significant increasing trends among arctic tundra buffers and significant decreasing trends among interior boreal forest buffers. The interannual variability of NDVI among the arctic tundra buffers was related to the previous summer warmth index. The spatial pattern of increasing tundra NDVI at the pixel level was related to the west-to-east spatial pattern in changing climate across arctic Alaska. There was no significant relationship between interannual NDVI and precipitation or temperature among the boreal forest buffers. The decreasing NDVI trend in interior boreal forests may be due to several factors including increased insect/disease infestations, reduced photosynthesis and a change in root/leaf carbon allocation in response to warmer and drier growing season climate. Main conclusions There was a contrast in trends of 1982,2003 annual maximum NDVI, with cold arctic tundra significantly increasing in NDVI and relatively warm and dry interior boreal forest areas consistently decreasing in NDVI. The annual maximum NDVI from arctic tundra areas was strongly related to a summer warmth index, while there were no significant relationships in boreal areas between annual maximum NDVI and precipitation or temperature. Annual maximum NDVI was not related to spring NDVI in either arctic tundra or boreal buffers. [source]

Impact of spatial variations of land surface parameters on regional evaporation: a case study with remote sensing data

Hussein O. Farah
Abstract Most precipitation in watersheds is consumed by evaporation, thus techniques to appraise regional evaporation are important to assess the availability of water resources. Many algorithms to estimate evaporation from remotely sensed spectral data have been developed in the recent past. In addition to differences in the physical parameterization of surface fluxes, these algorithms have different solutions for describing spatial variations of the parameters in the soil,vegetation,atmosphere,transfer (SVAT) continuum. In this study, the necessity to spatially distinguish SVAT parameters for computing surface heat fluxes is analysed for the Naivasha watershed in the Kenyan Rift Valley. Landsat Thematic Mapper (TM) spectral data have been used to first delineate the watershed into 15 hydrological units using surface temperature, normalized difference vegetation index and surface albedo as attributes. Thereafter, semi-empirical relationships between these TM-based parameters and other SVAT parameters have been applied to compute the spatial variation of SVAT parameters and the associated evaporation from the different hydrological units. The impact of using watershed-constant or watershed-distributed SVAT parameters on the fluxes is analysed. The determination of watershed averaged evaporation with area-aggregated SVAT parameters is feasible without significant loss of accuracy. Distributed evaporation in heterogeneous watersheds, however, can be investigated only with remote sensing flux algorithms that can account for spatially variable air temperature, surface roughness, surface albedo and the stability correction of the temperature profile due to buoyancy. Erroneous results can be expected if area-aggregated SVAT parameters are used to calculate local evaporation. As most of the recently developed remote sensing flux algorithms are based on areal constant SVAT parameters, direct applications in watersheds are still limited. Copyright © 2001 John Wiley & Sons, Ltd. [source]

Global analyses of satellite-derived vegetation index related to climatological wetness and warmth

Rikie Suzuki
Abstract Wetness and warmth are the principal factors that control global vegetation distribution. This paper investigates climate,vegetation relationships at a global scale using the normalized difference vegetation index (NDVI), warmth index (WAI), and wetness index (WEI). The NDVI was derived from a global, 20-year Advanced Very High Resolution Radiometer (AVHRR) dataset with 4-min resolution. The WEI was defined as the ratio of precipitation to potential evaporation. The WAI was defined as the cumulative monthly mean temperature that exceeds 5 °C annually. Meteorological data from the International Satellite Land-Surface Climatology Project Initiative II (ISLSCP II) dataset were used to calculate the WEI and WAI. All analyses used annual values based on averages from 1986 to 1995 at 1 × 1 degree resolution over land. Relationships among NDVI, WEI, and WAI values were examined using a vegetation-climate diagram with the WEI and WAI as orthogonal coordinates. The diagram shows that large NDVI values correspond to areas of tropical and temperate forests and large WEI and WAI values. Small WEI and WAI values are associated with small NDVI values that correspond to desert and tundra, respectively. Two major regimes are revealed by the NDVI vegetation-climate diagram: wetness dominant and warmth dominant. Wetness dominates mid- and low latitudes. Warmth dominates high latitudes north of 60°N or elevated land such as the Tibetan Plateau. The boundary between the two regimes roughly corresponds to the vegetation boundary between taiga forest and southern vegetation. Over northern Eurasia, the boundary occurs in areas where the NDVI is large and the maximum monthly temperature is around 18 °C. Copyright © 2006 Royal Meteorological Society. [source]

Managing heterogeneity in elephant distribution: interactions between elephant population density and surface-water availability

Summary 1Concerns over the ecological impacts of high African elephant Loxodonta africana densities suggest that it may be necessary to control their numbers locally, although the best management approach is still widely debated. Artificial water supply is believed to be a major cause of local overabundance, and could be used as a potential tool to regulate elephant distribution and impact across landscapes, but its effect on elephants at the population scale has never been studied. 2We assessed how dry-season surface-water availability constrained the distribution of an entire elephant population, using aerial and waterhole census data from Hwange National Park, Zimbabwe. The study was initiated in 1986, when the population was released from culling. We studied how artificial waterholes, holding water throughout the dry season, and vegetation production, estimated from a normalized difference vegetation index (NDVI), influenced the long-term distribution of elephant densities. We also investigated how the elephant distribution responded to changes in population density and annual rainfall, a driver of surface-water availability. 3Long-term dry-season elephant densities across the park tended to increase with vegetation production, and increased asymptotically with the density of artificial waterholes. 4Since the culling stopped, dry-season elephant densities have increased in most areas of the park, except in areas of low vegetation production and low water availability. Interannual fluctuations in elephant distribution are linked to rainfall variability through its effect on surface-water availability. During dry years elephants concentrated in areas where artificial pumping maintained surface-water availability during the dry season. 5During dry years elephant numbers at waterholes increased because of reduced surface-water availability, and elephants were distributed more evenly across waterholes, although active waterholes were unevenly distributed across the park. 6Synthesis and applications. Surface-water availability drives the distribution and abundance of elephants within Hwange National Park, and therefore appears to be at the heart of the trade-off between elephant conservation and the extent of their impact on ecosystems. Artificial manipulation of surface water is one of the tools available for the management of elephant populations and should not be overlooked when considering options for controlling elephant numbers in places where they are considered to be overabundant. [source]

Patterns of density, diversity, and the distribution of migratory strategies in the Russian boreal forest avifauna

Russell Greenberg
Abstract Aim, Comparisons of the biotas in the Palaearctic and Nearctic have focused on limited portions of the two regions. The purpose of this study was to assess the geographic pattern in the abundance, species richness, and importance of different migration patterns of the boreal forest avifauna of Eurasia from Europe to East Asia as well as their relationship to climate and forest productivity. We further examine data from two widely separated sites in the New World to see how these conform to the patterns found in the Eurasian system. Location, Boreal forest sites in Russia and Canada. Methods, Point counts were conducted in two to four boreal forest habitats at each of 14 sites in the Russian boreal forest from near to the Finnish border to the Far East, as well as at two sites in boreal Canada. We examined the abundance and species richness of all birds, and specific migratory classes, against four gradients (climate, primary productivity, latitude, and longitude). We tested for spatial autocorrelation in both dependent and independent variables using Moran's I to develop spatial correlograms. For each migratory class we used maximum likelihood to fit models, first assuming uncorrelated residuals and then assuming spatially autocorrelated residuals. For models assuming unstructured residuals we again generated correlograms on model residuals to determine whether model fitting removed spatial autocorrelation. Models were compared using Akaike's information criterion, adjusted for small sample size. Results, Overall abundance was highest at the eastern and western extremes of the survey region and lowest at the continent centre, whereas the abundance of tropical and short-distance migrants displayed an east,west gradient, with tropical migrants increasing in abundance in the east (and south), and short-distance migrants in the west. Although overall species richness showed no geographic pattern, richness within migratory classes showed patterns weaker than, but similar to, their abundance patterns described above. Overall abundance was correlated with climate variables that relate to continentality. The abundances of birds within different migration strategies were correlated with a second climatic gradient , increasing precipitation from west to east. Models using descriptors of location generally had greater explanatory value for the abundance and species-richness response variables than did those based on climate data and the normalized difference vegetation index (NDVI). Main conclusions, The distribution patterns for migrant types were related to both climatic and locational variables, and thus the patterns could be explained by either climatic regime or the accessibility of winter habitats, both historically and currently. Non-boreal wintering habitat is more accessible from both the western and eastern ends than from the centre of the boreal forest belt, but the tropics are most accessible from the eastern end of the Palaearctic boreal zone, in terms of distance and the absence of geographical barriers. Based on comparisons with Canadian sites, we recommend that future comparative studies between Palaearctic and Nearctic faunas be focused more on Siberia and the Russian Far East, as well as on central and western Canada. [source]

Breeding bird species richness in Taiwan: distribution on gradients of elevation, primary productivity and urbanization

Pei-Fen Lee
Abstract Aim, To examine the richness of breeding bird species in relation to elevation, primary productivity and urbanization. Location, The island of Taiwan (120°,122° E, 22°,25° N). Methods, We arranged bird species richness (BSR) data from 288 bird censuses undertaken in Taiwan into a 2 × 2 km quadrat system and calculated average values of elevation, primary productivity [surrogated by normalized difference vegetation index (NDVI)], and urbanization (surrogated by road density and percentage of built area) for each 2 × 2 km quadrat. Results, Bird species richness showed a hump-shaped relationship with elevation. It increased with elevation from sea level (10,64 species per 2 × 2 km quadrat), peaked around 2000 m (43,76 species), and then decreased with elevation towards its minimum at the highest elevation. Road density and percentage of built area decreased with elevation, and NDVI showed a hump-shaped relationship with elevation and inverse relationships with road density and percentage of built area. BSR increased with NDVI and decreased with road density and percentage of built area. Linear and cubic terms of elevation together explained 31.3% of the variance in BSR, and road density explained additional 3.4%. The explanatory power of NDVI on BSR was insignificant after the effects of elevation and road density had been justified. Main conclusions, We argue that urbanization plays an important role in the BSR of Taiwan. Urbanization might indirectly decrease BSR through decreasing primary productivity and therefore change the hypothetical inverse relationship between BSR and elevation into a hump-shaped relationship. We also propose a time hypothesis that the biotic communities in the mid-elevation zone of Taiwan had relatively longer periods of existence during the Pleistocene glacial cycles, which might be one underlying process of the observed hump-shaped relationship between species diversity and elevation. [source]

Simulating the East African wildebeest migration patterns using GIS and remote sensing

Douglas E. Musiega
Abstract The Serengeti,Mara ecosystem in East Africa is a spectacular natural heritage endowed with diverse fauna and flora. The presence of the seasonally migrating wildebeest (Connochaetes taurinus) is a major boost for tourism. This migration however has enormous impacts to the ecosystem. Consequently efforts at monitoring the herd's migration trends and patterns remain a challenge to wildlife managers and ecologists in the region. In this paper, the relative influence of vegetation (normalized difference vegetation index), landscape and relief on herds migration routes are investigated and the migration routes simulated using GIS and remote sensing techniques. The results are compared with the annual mean route taken by the herds, as determined by radio tracking over the 1995,1997 period. Green vegetation availability is shown to be the major criterion in route choice. It is also shown that during the dry season phases of the migration (western trek, western corridor), the herd endures complex relief (complexity quantified based on slope and inter-visibility) in the search for greener grass. During the season of abundance (southern trek), relief becomes critical in making route choices, with herds avoiding difficult terrain, notwithstanding their relatively more abundant vegetation. The method proposed in this paper is viable for rapid prediction of approximate routes for the migrating wildebeest in different climatic conditions. Résumé L'écosystème Serengeti,Mara en Afrique Occidental est un patrimoine naturel spectaculaire, doté des divers variétés de flore et de faune. La présence du gnou migrateur (Connochaetes taurinus) représente un atout majeur pour le tourisme. Néanmoins, cette migration a un impact énorme sur l'écosystème. Par conséquence, la surveillance des tendances migratoires du troupeau est un défi constant pour les gérants et les écologistes dans la région. Dans cette enquête, l'influence relative de la végétation (NDVI), le paysage et le relief, sur les routes du migration prises par le troupeau ont étéétudiés, et simulés utilisant le Système d'Information Géographique (SIG) et des techniques de perception à distance. Les résultats sont comparés à la moyenne annuelle des routes prises par les troupeaux, déterminée par le repérage radio pendant la période allant de 1995 a 1997. Le disponibilité de végétation verte s'avère le déterminant majeur dans le choix du chemin. Il est aussi démontré que pendant les phases du migration en saison sèche (périple vers l'ouest, couloir vers l'ouest) le troupeau subit des reliefs complexes (complexité calculée sur la pente et inter visibilité) à la recherche de l'herbe la plus verte. Pendant la saison d'abondance (périple vers le sud), l'impact du relief sur le choix des routes devient critique, les troupeaux évitant le terrain difficile, malgré sa végétation relativement abondante. La méthode présentée dans cette étude permet de prédire de façon rapide et valable la route approximative des gnous en cours de migration dans des conditions climatiques diverses. [source]

Drivers of megaherbivore demographic fluctuations: inference from elephants

M. J. Trimble
Abstract Environmentally induced variation in survival and fecundity generates demographic fluctuations that affect population growth rate. However, a general pattern of the comparative influence of variation in fecundity and juvenile survival on elephant population dynamics has not been investigated at a broad scale. We evaluated the relative importance of conception, gestation, first year survival and subsequent survivorship for controlling demographic variation by exploring the relationship between past environmental conditions determined by integrated normalized difference vegetation index (INDVI) and the shape of age distributions at 17 sites across Africa. We showed that, generally, INDVI during gestation best explained anomalies in age structure. However, in areas with low mean annual rainfall, INDVI during the first year of life was critical. The results challenge Eberhardt's paradigm for population analysis that suggests that populations respond to limited resource availability through a sequential decrease in juvenile survival, reproductive rate and adult survival. Contrastingly, elephants appear to respond first through a reduction in reproductive rate. We conclude that this discrepancy is likely due to the evolutionary significance of extremely large body size , an adaptation that increases survival rate but decreases reproductive potential. Other megaherbivores may respond similarly to resource limitation due to similarities in population dynamics. Knowing how vital rates vary with changing environmental conditions will permit better forecasts of the trajectories of megaherbivore populations. [source]

The tick Ixodes ricinus: distribution and climate preferences in the western Palaearctic

Abstract In this study, multivariate spatial clustering on monthly normalized difference vegetation index (NDVI) maps is used to classify ecological regions over the western Palaearctic. This classification is then used to delineate the distribution and climate preferences of populations (clades) of the tick Ixodes ricinus L. (Acari: Ixodidae) from a geographically extensive dataset of tick records and a gridded 2.5-km resolution climate dataset. Using monthly layers of the NDVI, regions of similar ecological attributes were defined and nine populations with significant differences in critical climate parameters (P < 0.005) were detected. Grouping of tick records according to other categories, such as political divisions, a 4°× 4° grid overlying the study area, or the CORINE) and USGS) vegetation classification schemes did not provided significantly separated populations (P= 0.094,0.304). Factor analysis and hierarchical tree clustering provided an ecological overview of these tick clades: two Mediterranean and one Scandinavian (western) clades are clearly separated from a node that includes clades of different parts of central Europe and the British Isles, with contrasting affinities between the different clades. The capture records of these ecologically separated clades produce a clear bias when bioclimate envelope modelling is applied to the mapping of habitat suitability for the tick in the western Palaearctic. The best-performing methods (Cohen's kappa = 0.834,0.912) use partial models developed with data from each ecoregion, which are then overlapped over the region of study. It is concluded that the use of ecologically derived ecoregions is an objective step in assessing the presence of ecologically different clades, and provides a guide in the development of data partitioning for habitat suitability modelling. [source]

Modelling the distributions of Culicoides bluetongue virus vectors in Sicily in relation to satellite-derived climate variables

B. V. Purse
Abstract., Surveillance data from 268 sites in Sicily are used to develop climatic models for prediction of the distribution of the main European bluetongue virus (BTV) vector Culicoides imicola Kieffer (Diptera: Ceratopogonidae) and of potential novel vectors, Culicoides pulicaris Linnaeus, Culicoides obsoletus group Meigen and Culicoides newsteadi Austen. The models containing the ,best' climatic predictors of distribution for each species, were selected from combinations of 40 temporally Fourier-processed remotely sensed variables and altitude at a 1 km spatial resolution using discriminant analysis. Kappa values of around 0.6 for all species models indicated substantial levels of agreement between model predictions and observed data. Whilst the distributions of C. obsoletus group and C. newsteadi were predicted by temperature variables, those of C. pulicaris and C. imicola were determined mainly by normalized difference vegetation index (NDVI), a variable correlated with soil moisture and vegetation biomass and productivity. These models were used to predict species presence in unsampled pixels across Italy and for C. imicola across Europe and North Africa. The predicted continuous presence of C. pulicaris along the appenine mountains, from north to south Italy, suggests BTV transmission may be possible in a large proportion of this region and that seasonal transhumance (seasonal movement of livestock between upland and lowland pastures) even in C. imicola -free areas should not generally be considered safe. The predicted distribution of C. imicola distribution shows substantial agreement with observed surveillance data from Greece and Iberia (including the Balearics) and parts of mainland Italy (Lazio, Tuscany and areas of the Ionian coast) but is generally much more restricted than the observed distribution (in Sardinia, Corsica and Morocco). The low number of presence sites for C. imicola in Sicily meant that only a restricted range of potential C. imicola habitats were included in the training set and that predictions could only be made within this range. Future modelling exercises will use abundance data collected according to a standardized protocol across the Mediterranean and, for Sicily in particular, should include non-climatic environmental variables that may influence breeding site suitability such as soil type. [source]

Survival in a long-lived territorial migrant: effects of life-history traits and ecological conditions in wintering and breeding areas

OIKOS, Issue 4 2009
Juan M. Grande
Despite its key role in population dynamics and evolutionary ecology, little is known about factors shaping survival in long-lived territorial species. Here, we assessed several hypotheses that might explain variability in survival in a migratory Spanish population of a long-lived territorial species, the Egyptian vulture Neophron percnopterus, using a 16-year monitoring period and live-encounter histories of 835 individually marked birds. Cormack-Jolly-Seber capture,recapture models showed no evidence for effects of sex or nestling body condition on survival. However, the normalized difference vegetation index (NDVI; an indicator of primary productivity) of natal territories had positive effects on juvenile survival, indicating that environmental conditions experienced early in life can determine survival prospects. Survival increased with age (0.73±0.02 in the first 2 years to 0.78±0.03 in years 3 and 4) to later decrease when birds were five years old (0.60±0.05), the age at which they acquire the adult plumage, abandon the communal lifestyle of juveniles, and may look for a breeding territory. At older ages, survival was higher for non-breeding (0.75±0.02) and breeding adults (0.83±0.02). Among the latter, birds that recruited into better territories had higher survival prospects. Age-specific variation in survival in this species may be related to behavioural changes linked to dispersal and recruitment into the breeding population, while survival prospects of adult birds strongly depend on breeding territory selection. These results suggest a tradeoff between recruiting soon, and thus reducing mortality costs of a long and extensive dispersal period, and trying to recruit into a good quality territory. Finally, annual survival rates for birds of all age classes were positively related with the NDVI in their African wintering grounds. Although this relationship was probably mediated by food availability, further research is needed to properly identify the limiting factors that are affecting trans-Saharan migrants, especially in light of global climate change. [source]

Short-term propagation of rainfall perturbations on terrestrial ecosystems in central California

Mónica García
Abstract Question: Does vegetation buffer or amplify rainfall perturbations, and is it possible to forecast rainfall using mesoscale climatic signals? Location: Central California (USA). Methods: The risk of dry or wet rainfall events was evaluated using conditional probabilities of rainfall depending on El Niño Southern Oscillation (ENSO) events. The propagation of rainfall perturbations on vegetation was calculated using cross-correlations between monthly seasonally adjusted (SA) normalized difference vegetation index (NDVI) from the Advanced Very High Resolution Radiometer (AVHRR), and SA antecedent rainfall at different time-scales. Results: In this region, El Niño events are associated with higher than normal winter precipitation (probability of 73%). Opposite but more predictable effects are found for La Niña events (89% probability of dry events). Chaparral and evergreen forests showed the longest persistence of rainfall effects (0-8 months). Grasslands and wetlands showed low persistence (0-2 months), with wetlands dominated by non-stationary patterns. Within the region, the NDVI spatial patterns associated with higher (lower) rainfall are homogeneous (heterogeneous), with the exception of evergreen forests. Conclusions: Knowledge of the time-scale of lagged effects of the non-seasonal component of rainfall on vegetation greenness, and the risk of winter rainfall anomalies lays the foundation for developing a forecasting model for vegetation greenness. Our results also suggest greater competitive advantage for perennial vegetation in response to potential rainfall increases in the region associated with climate change predictions, provided that the soil allows storing extra rainfall. [source]

Spatial variability of above-ground net primary production in Uruguayan grasslands: a remote sensing approach

S. Baeza
Abstract Question: How does above-ground net primary production (ANPP) differ (estimated from remotely sensed data) among vegetation units in sub-humid temperate grasslands? Location: Centre-north Uruguay. Methods: A vegetation map of the study area was generated from LANDSAT imagery and the landscape configuration described. The functional heterogeneity of mapping units was analysed in terms of the fraction of photosynthetically active radiation absorbed by green vegetation (fPAR), calculated from the normalized difference vegetation index (NDVI) images provided by the moderate resolution imaging spectroradiometer (MODIS) sensor. Finally, the ANPP of each grassland class was estimated using NDVI and climatic data. Results: Supervised classification presented a good overall accuracy and moderate to good average accuracy for grassland classes. Meso-xerophytic grasslands occupied 45% of the area, Meso-hydrophytic grasslands 43% and Lithophytic steppes 6%. The landscape was shaped by a matrix of large, unfragmented patches of Meso-xerophytic and Meso-hydrophytic grasslands. The region presented the lowest anthropic fragmentation degree reported for the Rio de la Plata grasslands. All grassland units showed bimodal annual fPAR seasonality, with spring and autumn peaks. Meso-hydrophytic grasslands showed a radiation interception 10% higher than the other units. On an annual basis, Meso-hydrophytic grasslands produced 3800 kg dry matter (DM) ha,1 yr,1 and Meso-xerophytic grasslands and Lithophytic steppes around 3400 kg·DM·ha,1·yr,1. Meso-xerophytic grasslands had the largest spatial variation during most of the year. The ANPP temporal variation was higher than the fPAR variability. Conclusions: Our results provide valuable information for grazing management (identifying spatial and temporal variations of ANPP) and grassland conservation (identifying the spatial distribution of vegetation units). [source]

Assessing species density and abundance of tropical trees from remotely sensed data and geostatistics

J. Luis Hernández-Stefanoni
Abstract Question: What relationships exist between remotely sensed measurements and field observations of species density and abundance of tree species? Can these relationships and spatial interpolation approaches be used to improve the accuracy of prediction of species density and abundance of tree species? Location: Quintana Roo, Yucatan peninsula, Mexico. Methods: Spatial prediction of species density and abundance of species for three functional groups was performed using regression kriging, which considers the linear relationship between dependent and explanatory variables, as well as the spatial dependence of the observations. These relationships were explored using regression analysis with species density and abundance of species of three functional groups as dependent variables, and reflectance values of spectral bands, computed NDVI (normalized difference vegetation index), standard deviation of NDVI and texture measurements of Landsat 7 Thematic Mapper (TM) imagery as explanatory variables. Akaike information criterion was employed to select a set of candidate models and calculate model-averaged parameters. Variogram analysis was used to analyze the spatial structure of the residuals of the linear regressions. Results: Species density of trees was related to reflectance values of TM4, NDVI and spatial heterogeneity of land cover types, while the abundance of species in functional groups showed different patterns of association with remotely sensed data. Models that accounted for spatial autocorrelation improved the accuracy of estimates in all cases. Conclusions: Our approach can substantially increase the accuracy of the spatial estimates of species richness and abundance of tropical tree species and can help guide and evaluate tropical forest management and conservation. [source]

Grassland productivity in an alpine environment in response to climate change

AREA, Issue 3 2005
Yong Zha
Situated in a climatically stressful environment, alpine grassland is sensitive to subtle climate changes in its productivity. We remedy the current deficiency in studying grassland productivity by taking the integrated effect of all relevant factors into consideration. The relative importance of temperature, rainfall and evaporation to the alpine grassland productivity in western China was determined through analysis of their relationship with the normalized difference vegetation index (NDVI) between 1981 and 2000. Climate warming stimulated grassland productivity in the 1980s, but hampered it in the 1990s. Temperature is more important than rainfall to grassland productivity early in the growing season. However, their relative importance is reversed late in the growing season. Monthly summer month rainfall modified by maximum monthly temperature is a good predictor of alpine grassland productivity at 62.0 per cent. However, the best predictor is water deficiency, which is able to improve the estimation accuracy to 78.3 per cent. Hence, the impact of temperature on grassland productivity is better studied indirectly through evaporation. [source]