Informations et ressources scientifiques
sur le développement des zones arides et semi-arides

Accueil du site → Master → Etats Unis → 2022 → Satellite Estimates of Grass Productivity in a Heterogenous Southwestern Rangeland

New Mexico State University (2022)

Satellite Estimates of Grass Productivity in a Heterogenous Southwestern Rangeland

Hansen, Taylor M

Titre : Satellite Estimates of Grass Productivity in a Heterogenous Southwestern Rangeland

Auteur : Hansen, Taylor M

Université de soutenance : New Mexico State University

Grade : Master of Science (MS) 2022

Description
Drylands cover approximately 40% of the Earth’s land surface but there are relatively few remotely sensed vegetation indices that are tailored to the challenges that come with observing arid and semi-arid rangelands from space. Low and heterogeneous vegetation cover and exposed soils along with a mixture of woody and herbaceous cover present problems for discerning forage abundance and distribution. A new vegetation index, the Normalized Differential Phenometric Index (NDPI), was developed to overcome some of these challenges by taking advantage of the different phenology types typical in dryland landscapes while also minimizing the effects of soils. Shrubs and grasses in particular have different phenology and react to climate drivers differently. Here I test the ability of the NDPI to capture the fraction of grass and shrub cover at the Jornada Experimental Range (JER) in Southern New Mexico and examine the ability of the NDPI to represent the spatial variability of precipitation impacts on grass competition with shrub cover. MODIS/Terra surface reflectance 8-day image composites with a spatial resolution of 250 m were used to test the ability of NDPI to separate grass and shrub features at the Jornada. I calculated NDPI over 15 years, from 2005-2019 using Google Earth Engine. The results show that the NDPI approach provides for the initial detection of shrubs, where shrub cover is correlated with a late dry season vegetation index (VI). This allows monsoon season C4 grass growth to be monitored as the difference between wet and dry season VI. The results also show that grass productivity is correlated with seasonal rainfall, as expected, but also indicate that shrubs may play a more significant role in competition for resources in wetter years. To the best of my knowledge this is the first time that interaction has been identified at regional scale using remotely sensed data. In addition, I found that while rainfall from the coarse scale climate analysis product PRISM provides reasonably accurate average rainfall estimates, comparison to local (30-meter resolution) estimates made by interpolation between rain gauges in the Jornada meteorological station network shows that local rainfall heterogeneity is poorly represented in the PRISM data.

Présentation (NMSU )

Page publiée le 13 décembre 2022