In our ‘Field Diaries’ series, The Applied Ecologist is sharing stories from a range of different fieldwork experiences. In this post, Mohammad Javad Soltani shares their story visiting wetlands in Iran and studying them with remote sensing.
Overview of fieldwork project
I am a M.Sc. student in the ‘Remote Sensing for Ecology and Ecosystem Conservation (RSEEC)’ lab at the K. N. Toosi University of Technology, led by my supervisor Prof. Hooman Latifi. In my research, I try to illustrate the huge damage that unstructured and aggressive agricultural activities have imposed on Bisheh Dalan, a vital yet degraded wetland habitat in western Iranian province of Lorestan, using satellite imagery.
In 2024, I visited Bisheh Dalan to collect field data on wetland water bodies and other land cover-land use classes. Some striking key observations I made during the field visit, and some tips and tricks for fieldwork include:
- The wetland has become seasonal during the last decade, so it is crucial to study time series satellite imagery to understand the dynamics of both the water body and the vegetation. This also unveils the crucial timeslots for any subsequent field visits.
Video Link: 50-year satellite image timelapse
- I leveraged freely available satellite data and collected cloud free images during 1984-2024 using Google Earth Engine. In this step, understanding data gaps is really important! As you can see below, I observed some gaps in Landsat-5 image collection:
- After setting a proper boundary for your work, it is time for planning the field data collection, a very important part of the job. It is obvious that everything is going to be different in the field and unpredicted observations are expected to happen out there, but we cannot just neglect this by going out for a blind sampling! I visited Bisheh Dalan twice (a one-day visit in Early spring to assess the largest water body and a one-day visit in early summer) and performed a preliminary image-based analysis by using satellite timelapses before planning my main field data collection. So, I had already gained quite a comprehensive understanding! First, I calculated the well-known Normalized Difference Vegetation Index (NDVI) and Modified Normalized Difference Water Index (MNDWI), for the median images of 2023 and 2024 (below). NDVI is the most widely used index to extract vegetation, with temporal hotspots that can be extremely helpful to initialize field sampling, specially across those places in which NDVI presents changes throughout consecutive years.
- A further useful product to ease the fieldwork across complex wetland ecosystems can be constructing RGB colour composites from NDVI in three different times of the year. For instance, 90% max NDVI in March, July, and September can be helpful (see below). Can you see the true power of time series data? This incredible image was born by just using a single index at three timeslots! Using this and the previous step, you might be able to set your initial plan and data collection path by gaining a prior knowledge of different crop types prevailing the study site.
- A simple yet crucial step is definitely to check the weather forecast prior to field sampling over mixed wetland ecosystems. This is of great significance if you are particularly working with synthetic aperture radar (SAR) data, as even a one-day rainfall prior to the field sampling can notably affect the landcover spectral properties and signatures due to the height impact of Di-Electric constant on the recorded SAR signals.
- If you are working with optical satellite sensor data like Sentinel-2 or Landsat, there is no need to use multi-frequency GNSS and maintain a geometric accuracy in decimetre level, due to relatively coarse, i.e. 10m, spatial resolution of these sensors. In such cases, user-friendly mobile apps like “GPS-WAYPOINTS” can be considered a great asset along with Google Earth Pro snapshots or timelapses to process your raw data. I highly recommend you to use the temporally closest Sentinel-2 image to your field data collection besides Google Earth Pro, and check these two data for each sample that you collect in the field. Also, please try to keep the accuracy of the mobile app, e.g. GPS WAYPOINTS, constantly under 2-3 meters, and use 1 second sampling rate to collect data while you’re moving in the field.
- As I generally recommend collecting coordinates of paths instead of points, another practical idea is to record a video snapshot from a given sample, store the data with a unique ID, and bear in mind that it would alleviate further processing of this sample in the lab posterior to the data sampling. These movies will come handy in conditions where sample locations or properties could be mixed-up due to their high number or similar signatures! Thus, take it seriously!
- [Highly important!] You must definitely take time to communicate with local administrative bodies like, in my case, the protected area rangers who are responsible for ecosystem protection or monitoring. This can also be extended to literally anyone who is familiar with the surrounding environment. Further, the risk of wild animals, hunters, or swamps as risk factors cannot be overlooked! You must avoid risky areas as far as possible. However, if it is absolutely necessary to collect data from those places, using unmanned aerial vehicles (UAVs) can offer a pragmatic alternative! During my own field sampling, the risk was getting too close to wolves and wild boars habitats.
Video Link: Wolves and wild boars – fieldwork insights
- Try to avoid accessing habitats of animals and birds in the area or minimize to the least possible circumstances, in particular if your research is not directly related to animal ecology. In case of encounters, do not feed them, as you might [incautiously] disrupt the food chain! Also make sure to pack all your equipment and garbage at the end of your fieldwork and, again, in particular case of water-dominated ecosystems like wetlands.
- Be patient with locals and engage with them in your fieldwork, at least by informing them on the objectives and potential assets of your research work. They may provide valuable insights, such as the wetland’s history or recent weather patterns.
Why is this project important?
Practicing intensive agriculture inside and around wetland ecosystems poses serious ecological challenges at regional levels and beyond. In semi-arid countries like Iran, the effects can be exacerbated as a result of droughts raised by climate change, which not only threatens the sustainability of wetlands, but also local livelihoods in the mid- and long-term. Tracking these trends is only possible by leveraging time series of remote sensing data coupled with advanced pattern recognition techniques.
In my case, even a simple trend extracted from the NDVI data (above), is helpful in unveiling the phenology of each vegetation and wetland class, where a retrospective monitoring of these classes might lead to detecting the effects of land use conversion by intensive cultivation on degrading wetlands.
Next steps
One of the most powerful tools in modern conservation is the use of satellite remote sensing data, particularly in the context of time series analysis. This enables both researchers and policymakers to track changes over time, identifying critical “hotspots” of ecological transformation or detecting emerging trends. For example, with remote sensing, we can monitor the shifting water body of wetlands, as well as the health and distribution of key wetland vegetation species like Phragmites sp. or Tamarix sp.
What makes this approach innovative is the ability to cross-track these natural changes with agricultural land use. By analyzing crop patterns alongside wetland ecosystems, remote sensing allows detection of correlations, such as how specific crops may impact water levels or compete with native vegetation. This cross-comparison provides essential insights that can guide sustainable land management strategies, helping us safeguard wetlands while optimizing agricultural productivity.
Another crucial area of ongoing research in wetland ecosystems is the ability to assess the economic value of ecosystem services. By doing so, we can integrate this information into long-term conservation planning, ensuring that wetlands sustainably maintain their ecosystem services to both wildlife and humans. This could not only restore wetlands but also establish clear, sustainable boundaries for preservation. In the long term, remote sensing, combined with dynamic mapping, can play a vital role in this process. Creating detailed Land Use and Land Cover (LULC) maps will allow tracking changes in the wetland and its surroundings, providing the data needed for adaptive management strategies. These approaches can help balance ecological conservation with the needs of local communities, ensuring the wetland’s health and resilience for future generations.
Discover more posts from our ‘Fieldwork Diaries‘ blog series here. If you have an idea for a blog that would fit into this series, please contact Catherine Waite.
