Riparian Area Changes in Greenness and Water Use on the Lower Colorado River in the USA from 2000 to 2020

Pamela L Nagler,Christopher J Jarchow,Sattar Chavoshi Borujeni,Kamel Didan,Hamideh Nouri,Armando Barreto-Muñoz

doi:10.3390/rs13071332

Abstract

Declines in riparian ecosystem greenness and water use have been observed in the delta of the Lower Colorado River (LCR) since 2000. The purpose of our case study was to measure these metrics on the U.S. side of the border between Hoover and Morelos Dams to see if declining greenness was unique to the portion of the river in Mexico. In this case study, five riparian reaches of the LCR from Hoover to Morelos Dam since 2000 were studied to evaluate trends in riparian ecosystem health. We measure these riparian woodlands using remotely sensed measurements of the two-band Enhanced Vegetation Index (EVI2; a proxy for greenness); daily evapotranspiration (ET; mmd−1) using EVI2 (ET(EVI2)); and an annualized ET based on EVI2, the Phenology Assessment Metric (PAM ET), an annualized ET using Landsat time-series. A key finding is that riparian health and its water use has been in decline since 2000 on the U.S. portion of the LCR, depicting a loss of green vegetation over the last two decades. EVI2 results show a decline of −13.83%, while average daily ET(EVI2) between the first and last decade had a decrease of over 1 mmd−1 (−27.30%) and the respective average PAM ET losses were 170.91 mmyr−1 (−17.95%). The difference between the first and last five-year periods, 2000–2005 and 2016–2020, showed the largest decrease in daily ET(EVI) of 1.24 mmd−1 (−32.61%). These declines come from a loss in healthy, green, riparian plant-cover, not a change in plant water use efficiency nor efficient use of managed water resources. Our results suggest further deterioration of biodiversity, wildlife habitat and other key ecosystem services on the U.S. portion of the LCR.

Highlights

We assessed Normalized Difference Vegetation Index (NDVI) and NDVI*, but our research results are primarily confined to data derived from EVI2: EVI2 for the peak growing season, EVI2 for the full year, ET with EVI2 as the input vegetation index (VI), and Phenology Assessment Metric (PAM ET) using
We produced time-series data for three VIs including NDVI, EVI, and EVI2, but because we found that EVI and EVI2 had an r2 = 0.99, we proceeded with using only NDVI
NDVI (Figure 3a) and EVI2 (Figure 3b) were produced following our methods which account for amplitude shifts across three sensors, Landsat5 (TM5), Landsat7 (ETM+), and Landsat8 (OLI)

Summary

Introduction

Accurate estimates of evapotranspiration (ET; water evaporated by bare soil and transpired by riparian plants) for riparian, urban green spaces, and cultivated lands are needed for management tasks at all scales. These tasks include scheduling for irrigation; sustaining agricultural production; securing foods and safe water quality and quantity for human uses; managing watersheds; allocating water; determining water rights; forecasting weather; and monitoring, managing, and projecting the long-term effects of land use change and global climate change on water resources [1,2,3,4,5]. Measured and validated riparian ET [25,26,27,28,29,30] and urban ET [31] have rarely been included in water budgets because of their small land cover relative to vast agricultural plots and, because of this, the majority of remotely sensed ET work is focused on agricultural water balance methods and models (i.e., crop coefficients, energy balance, and thermal methods) [32,33,34,35,36,37]

Objectives

Methods

Results

Discussion

Conclusion