Evaluating the Relationships of Inter-Annual Farmland Vegetation Dynamics with Biodiversity Using Multi-Spatial and Multi-Temporal Remote Sensing Data

Niloofar Alavi,Douglas King

doi:10.3390/rs12091479

Niloofar Alavi, Douglas King

Open Access

https://doi.org/10.3390/rs12091479

Copy DOI

Abstract

Agricultural landscapes are highly dynamic ecosystems, but the effects of temporal farmland vegetation dynamics on species diversity have not been widely studied. In 93 sample farm landscapes in eastern Ontario, Canada, biodiversity data for seven taxa were collected in 2011 and 2012, prior to the initiation of this study. The goal of this study was to determine if trends and variability in vegetation productivity detected in these sample landscapes using long-term archived moderate and coarse resolution remote sensing time series data are related to the measured biodiversity. Mid-summer Moderate Resolution Imaging Spectroradiometer (MODIS) (2000–2011) and Landsat 5 (1985–2011) Normalized Difference Vegetation Index (NDVI) data were used with the Thiel–Sen slope and Contextual Mann–Kendall trend analysis to identify pixels showing significant trends. NDVI temporal metrics included 1) the percentage of pixels in each landscape with a significant negative or positive trend, and 2) the temporal coefficient of variation (CV) of both the mean and spatial CV of landscape NDVI. Larger areas of significant positive NDVI trends were found in the sample landscapes than negative trends, the former being associated with agricultural intensification or crop changes and the latter with smaller areas of natural vegetation removal. Landsat better-detected changes in individual fields or small areas of natural vegetation due to its much smaller pixel size. In addition, the longer Landsat time series showed a change in the NDVI trend from positive (1985–2000) to negative or a leveling off (2000–2011) for many pixels. In biodiversity modeling, the Landsat temporal CV of NDVI was negatively correlated with 2011–2012 plant and beetle diversity, while plant biodiversity was positively correlated with the percentage of pixels in a sample landscape showing a significantly positive NDVI trend. No significant relationships were found using the MODIS data. This study shows that temporal trends and variability in farmland vegetation density derived from Landsat data are related to biodiversity for certain taxa and that such relationships should be considered along with the more commonly studied spatial landscape attributes in evaluating landscape-level impacts of farming on biodiversity.

Highlights

Biodiversity has declined significantly in recent decades, mainly due to habitat loss and habitat fragmentation [1]
The evaluation of the TS slope and Contextual Mann–Kendall (CMK) significance for 2000–2011 July average Moderate Resolution Imaging Spectroradiometer (MODIS) Normalized Difference Vegetation Index (NDVI) data within the sample landscapes shows that 38% of all pixels demonstrated a statistically significant trend in NDVI
It is evident that more sample landscapes had greater proportions of pixels with significant positive trends in NDVI than negative trends

Summary

Introduction

Biodiversity has declined significantly in recent decades, mainly due to habitat loss and habitat fragmentation [1]. The intensification of farming in existing agricultural lands presents increased risks to biodiversity and species at risk and negatively affects natural habitats and ecosystems [3]. The capacity of Canadian farmlands in carrying wildlife species was relatively stable from 1986 to 1996; it declined from 1996 to 2011, with agricultural intensification being a leading cause of this decrease [5]. Conservation and alternative agricultural practices are being adopted in some areas to revitalize the landscape, enhance biodiversity, and minimize damage to living organisms. Organic farming, crop rotation, no-tillage systems, and management of natural field margins are some practices that promote biodiversity [6,7,8]. At the landscape level, which is the focal scale of this study, increasing spatial heterogeneity is often suggested as a means to enhance species richness and abundance [9,10]

Objectives

Methods

Results

Discussion

Conclusion