Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

TerraTrace: Temporal Signature Land Use Mapping System

Angela Busheska, Vikram Iyer, Bruno Silva, Peder Olsen, Ranveer Chandra, Vaishnavi Ranganathan

TL;DR

TerraTrace tackles the challenge of distinguishing cropland from forests in satellite-based land-use maps by leveraging temporal NDVI signatures. The approach combines a newly constructed longitudinal NDVI dataset for California (2020–2023) at 500 m resolution with an end-to-end platform that analyzes NDVI trajectories and augments this with CDL data and an LLM-assisted interpretation module. Preliminary demonstrations show that crop growth cycles imprint distinctive NDVI curves and that these signals align with CropScape validation, suggesting potential for scalable, low-overhead historical land-use tracking. The work has practical implications for regulatory compliance, climate monitoring, and agricultural decision-making, with clear paths toward global expansion and integration of additional spectral indices.

Abstract

Understanding land use over time is critical to tracking events related to climate change, like deforestation. However, satellite-based remote sensing tools which are used for monitoring struggle to differentiate vegetation types in farms and orchards from forests. We observe that metrics such as the Normalized Difference Vegetation Index (NDVI), based on plant photosynthesis, have unique temporal signatures that reflect agricultural practices and seasonal cycles. We analyze yearly NDVI changes on 20 farms for 10 unique crops. Initial results show that NDVI curves are coherent with agricultural practices, are unique to each crop, consistent globally, and can differentiate farms from forests. We develop a novel longitudinal NDVI dataset for the state of California from 2020-2023 with 500~m resolution and over 70 million points. We use this to develop the TerraTrace platform, an end-to-end analytic tool that classifies land use using NDVI signatures and allows users to query the system through an LLM chatbot and graphical interface.

TerraTrace: Temporal Signature Land Use Mapping System

TL;DR

TerraTrace tackles the challenge of distinguishing cropland from forests in satellite-based land-use maps by leveraging temporal NDVI signatures. The approach combines a newly constructed longitudinal NDVI dataset for California (2020–2023) at 500 m resolution with an end-to-end platform that analyzes NDVI trajectories and augments this with CDL data and an LLM-assisted interpretation module. Preliminary demonstrations show that crop growth cycles imprint distinctive NDVI curves and that these signals align with CropScape validation, suggesting potential for scalable, low-overhead historical land-use tracking. The work has practical implications for regulatory compliance, climate monitoring, and agricultural decision-making, with clear paths toward global expansion and integration of additional spectral indices.

Abstract

Understanding land use over time is critical to tracking events related to climate change, like deforestation. However, satellite-based remote sensing tools which are used for monitoring struggle to differentiate vegetation types in farms and orchards from forests. We observe that metrics such as the Normalized Difference Vegetation Index (NDVI), based on plant photosynthesis, have unique temporal signatures that reflect agricultural practices and seasonal cycles. We analyze yearly NDVI changes on 20 farms for 10 unique crops. Initial results show that NDVI curves are coherent with agricultural practices, are unique to each crop, consistent globally, and can differentiate farms from forests. We develop a novel longitudinal NDVI dataset for the state of California from 2020-2023 with 500~m resolution and over 70 million points. We use this to develop the TerraTrace platform, an end-to-end analytic tool that classifies land use using NDVI signatures and allows users to query the system through an LLM chatbot and graphical interface.

Paper Structure

This paper contains 6 sections, 1 equation, 4 figures.

Table of Contents

  1. Introduction
  2. NDVI Crop Signature Curves
  3. NDVI Dataset Preprocessing
  4. TerraTrace Platform
  5. Initial Demonstration and Results
  6. Discussion and Future Work

Figures (4)

  • Figure 1: Land use classification challenges (A) Current forest probability maps based on prior datasets ALOS-paperpittman2019global cannot distinguish between farms and wild forests. (B) Unlike prior work prestowstatt, TerraTrace tracks historical land use with low computational overhead and variable scales.
  • Figure 2: NDVI Signature Curves. (A) Comparison of NDVI on a wheat farm versus adjacent pine trees in Farmington, WA from Apr-Dec 2020. (B) NDVI curves for a chickpea farm in Grand Rapids, ND, an apple orchard in Wenatchee, WA and a citrus farm in Fresno County, CA. (C) NDVI for deciduous forest in MO, USA versus evergreen forest in WA, USA. (D) NDVI data for coffee farms in Buon Ma Thuot, Vietnam and El Paraiso, Honduras. (E) NDVI dataset created for CA, USA.
  • Figure 3: TerraTrace System. TerraTrace finds the NDVI coordinates from our dataset, extracts a set of metrics to analyze the data, and passes these to GPT-4 Turbo for additional analysis.
  • Figure 4: Screenshots of TerraTrace