Regression models

Early approaches to assess primary production in grasslands, such as the work by H. N. Le Houerou in the 1980s, used empirical models which correlated primary productivity with mean annual precipitation or evapotranspiration. In general, the relationships of production with environmental variables were derived from long-term averages for many sites distributed across environmental gradients (spatial models). ANPP increases linearly along spatial precipitation gradients within the range of 200-1300 mm yr~ in North American, South American, and African grasslands.

However, much less is known about the controls of the temporal, inter-annual variation of productivity at a given site (temporal models). Temporal models relating time series of ANPP and annual precipitation for single sites have shown lower slopes (= water use efficiency) and regression coefficients than the spatial models. Additionally, memory and carryover effects, for example, due to storage of carbohydrates in the root system and structural inertia, might play an important role in the functioning of semiarid grasslands. Memory and carryover effects buffer fluctuations in production if wet, productive years alternate with dry, less productive years and amplify fluctuations if wet or dry sequences of several years take place. Identifying and quantifying such memory and carryover effects is an important challenge for global models which mostly use simple linear relationships with precipitation.

Use of remote sensing data

Biomass harvesting is the most common way to estimate ANPP in grasslands, but because of the large effort and the detailed spatial scale, harvesting methods are rather limited in their spatial and temporal extent. Remote-sensing techniques are a fast and nondestructive method for estimating ANPP at a regional scale over longer time periods. NDVI is the most commonly used radiometric index for estimating ANPP in grasslands and the annual summed NDVI can be used as a surrogate for annual ANPP because, in ecosystems dominated by grasses or deciduous life forms, the absorbed photosynthetically active radiation of plant canopies (APAR) and net primary production are directly related. Biomass estimates for grassland using NDVI have been performed, for example, by the group of J. M. Paruelo in the late 1990s and 2000s for the Central Grassland Region of the United States, and for subhumid pampa rangelands and the Patagonia steppes in Argentina. S. D. Prince determined in the early 1990s primary production for Sahelian grasslands.

Additionally, NDVI data allow determination of other important ecosystem characteristics such as the degree of seasonality, and the start and the end of the growing season. Relating NDVI characteristics over regional gradients with climatic and other environmental variables allow for inference on the controls of primary production and ecosystem functioning. Remote-sensing data may also be used to calibrate ecosystem models, for example, by minimizing the difference between the measured and the simulated NDVI. The utility of this approach was demonstrated by a study of Y. S. Nouvellon in the 2000s who coupled a grassland ecosystem model for semiarid perennial grasslands in southeastern Arizona (USA) with Landsat imagery for a 10-year simulation of carbon and water budgets.

Project Earth Conservation

Project Earth Conservation

Get All The Support And Guidance You Need To Be A Success At Helping Save The Earth. This Book Is One Of The Most Valuable Resources In The World When It Comes To How To Recycle to Create a Better Future for Our Children.

Get My Free Ebook


Post a comment