Relations between productivity, climate, and normalized difference vegetation index in the central Great Plains.

摘要

Understanding the influences of climate on productivity remains a major challenge in landscape ecology. Satellite remote sensing of normalized difference vegetation index (NDVI) provides a useful tool to study landscape patterns, based on generalization of local measurements, and to examine relations between climate and variation in productivity. This dissertation examines temporal and spatial relations between NDVI, productivity, and climatic factors over the course of nine years in the central Great Plains. Two general findings emerge: (1) integrated NDVI is a reliable measure of production, as validated with ground-based productivity measurements; and (2) precipitation is the primary factor that determines spatial and temporal patterns of NDVI.; NDVI, integrated over appropriate time intervals, is strongly correlated with ground productivity measurements in forests, grasslands, and croplands. Most tree productivity measurements (tree ring size, tree diameter growth, and seed production) are strongly correlated with NDVI integrated for a period during the early growing season; foliage production is most strongly correlated with NDVI integrated over the entire growing season; and tree height growth corresponds with NDVI integrate during the previous growing season. Similarly, productivity measurements for herbaceous plants (grassland biomass and crop yield) are strongly correlated with NDVI. Within the growing season, the temporal pattern of grassland biomass production covaries with NDVI, with a four-week lag time. Across years, grassland biomass production covaries with NDVI integrated from part to all of the current growing season. Corn and wheat yield are most strongly related to NDVI integrated from late June to early August and from late April to mid-May, respectively.; Precipitation strongly influences both temporal and spatial patterns of NDVI, while temperature influences NDVI only during the early and late growing season. In terms of temporal patterns, NDVI integrated over the growing season is strongly correlated with precipitation received during the current growing season plus the seven preceding months (fifteen month period); NDVI within the growing season responds to changes in precipitation with a four to eight week lag time; and major precipitation events lead to changes in NDVI with a two to four week lag time. Temperature has a positive correlation with NDVI during the early and late growing season, and a weak negative correlation during the middle of the growing season. In terms of spatial patterns, average precipitation is a strong predictor of the major east-west gradient of NDVI. Deviation from average precipitation explains most of the year-to-year variation in spatial patterns. NDVI and precipitation deviations from average covary (both positive or both negative) for 60–95% of the total land area in Kansas. Minimum and average temperatures are positively correlated with NDVI, but temperature deviation from average is generally not correlated with NDVI deviation from average.; The strong relationships between NDVI and productivity, and between precipitation and NDVI, along with detailed analysis of the temporal and spatial patterns for our study region, provides the basis for prediction of productivity at landscape scales under different climate regimes.