Setting Up a Workable Excavation Strategy
The first thing people get wrong about digging stuff up is that they treat it like treasure hunting. It isn't. You are sifting through decades, sometimes millennia, of layered accidental events. The theory side matters because without one, you are just moving dirt with extra steps. At the core, you need a framework. Processual archaeology treats the site as a system you can model and test. Post-processual work reminds you that human behavior is messy and context-dependent. Most field projects I have been on sit somewhere between the two. You start with a hypothesis about how the site formed, collect data systematically, and then see if your model holds up or needs revision. The cycle repeats until you run out of time and money. I worked a site in the Pacific Northwest a few years back where the standard stratigraphic approach kept falling apart. The ground had been bioturbated so badly by root systems and gopher burrows that the layers were scrambled over a two-meter depth. Standard grid excavation would have produced data that looked stratified but meant nothing. What I did instead was shift to a feature-based recording method. I stopped trying to define horizontal strata and started mapping individual intrusive features like postholes, hearths, and soil discoloration patches by their three-dimensional coordinates. It took longer on site but gave us recoverable context. The tradeoff is that you need a solid total station or RTK setup and people who know how to use them. If you are working on a shoestring with just tape measures, you are going to have a rough time with this approach.
Survey Before You Dig
Remote sensing and surface survey should always come before any shovel hits the ground. Ground penetrating radar can map subsurface anomalies quickly over a large area, but it struggles in wet clay and highly stony soils. Magnetometry works well for detecting hearths and fired features but gives noisy reads near modern metal debris. A combined approach is usually worth the extra cost and time. I typically budget about three to four days of survey for a site that ends up being roughly two hectares, depending on vegetation and terrain. Surface collection is another step people rush. Walking transects at twenty-meter intervals and picking up anything that looks modified will give you a baseline density map. You will miss small artifacts and things buried just under the litter layer, but it tells you where not to waste time and where the real concentration might be. A proper systematic surface collection with a five-meter interval and full sieve screening takes significantly longer but increases your artifact return by a noticeable margin. The question is whether your research design demands that level of detail.
Excavation Methodology
Unit layout depends on what you are looking for. Test grids work for initial assessment. Once you identify an area of interest, you switch to open-area excavation to expose larger surfaces. The choice between stratigraphic excavation and arbitrary levels is not a minor one. Stratigraphic excavation preserves sequence information but is slow. Arbitrary level excavation, usually at ten-centimeter intervals, is faster and can be sufficient when the deposit is homogenous or when you need volume quickly for flotation samples and bulk processing. Recording is where projects live or die. Every unit needs a context sheet, photography with scale and north arrow, and a section drawing. Digital photography replaced film for most teams, but you still need consistent lighting and color correction targets. I use a Datacolor ColorChecker in every shot now. It adds thirty seconds per photo and saves hours of post-processing headaches. Back in the day I skipped it and spent a full evening trying to white balance thirty photos taken under different cloud conditions. Never again. Soil removal should be done with trowels in most contexts. Power equipment has its place for clearing overburden, but the moment you hit cultural material you are on your own two hands. Screening is non-negotiable. Two-millimeter mesh catches fragments that are otherwise lost. For flotation, you want a finer recovery of botanical remains, which often turns out to be the most informative part of a site if you have the lab capacity to process it. One common mistake is screening everything through two millimeters and throwing away the oversize. Those oversized pieces are often diagnostic artifacts. Keep them separate.
Get the Full Details
Analysis and Interpretation
Artifact analysis is where specialization kicks in. You need someone who can identify ceramic types, lithic debitage patterns, faunal remains, and botanical samples. Trying to do all of it yourself as a single person usually means you end up doing none of it well. I have seen projects stall for months waiting on a single faunal report because the zooarchaeologist was backed up. Building your project timeline around realistic lab turnaround times is one of those things that sounds obvious until it bites you. Statistical analysis of spatial data has become standard. Point pattern analysis, kernel density estimation, and spatial autocorrelation can reveal activity areas that are not visible to the naked eye. The tools are accessible now. QGIS with the relevant plugins handles most of what you need without paying for proprietary software. The learning curve is real though. Expect a week of fiddling before you feel competent producing publishable maps. One counter-intuitive thing that trips people up is the assumption that more data is always better. A fully excavated site produces so much material that you may never process it all. Selective excavation of high-probability areas often yields a better research outcome than clearing everything. You preserve the rest for future methods that may be better than what you have today. That is not a philosophical stance. It is a practical one. Storage costs money, and repositories are getting full. There is no shame in leaving half a site in the ground.
Field Notes and Documentation Standards
Write things down as you go. Memory is unreliable even on good days. I keep a bound field notebook with grid references on every page and a separate logbook for context numbers. Digital tablets work in dry conditions but fail in rain and dust. Paper survives. My workflow is paper first, digital entry same evening if the battery and weather cooperate. The duplication is annoying but it has saved me more than once when a hard drive failed or a file corrupted. Photographic logs need to tie directly to your context numbers. I use a simple index sheet that links photo IDs to context numbers and unit locations. It takes ten minutes to set up and prevents the panic of not knowing which photo belongs to which layer. People who skip this step usually figure it out the hard way six months later when they are writing the report.
Reporting and Archiving
The final report is the product. Everything else is preparation for it. A good report includes methods, context summaries, artifact catalogs, analysis results, and discussion. The discussion section is where your theoretical framework actually matters. If you collected data without a question in mind, your report will read like an inventory. That is usable in a pinch but it does not advance anything. Make sure your research questions are stated clearly in the introduction and that your methods are designed to answer them. Archival deposition requirements vary by region and funding source. Some jurisdictions require the full artifact assemblage to be deposited at a museum within a year of the report. Others allow curation at a university repository. Check these requirements early. I once started a project assuming the artifacts would stay at my institution and ran into a compliance issue when the state historic preservation office demanded transfer. Resolving it cost about two weeks of administrative work that could have been avoided with a phone call at the planning stage. Lab processing timelines also depend on artifact type. Lithics can be analyzed relatively quickly by an experienced person. Ceramics take longer because typological classification is meticulous. Bone requires cleaning, sorting, and measurement. Botanical samples from flotation need drying and sorting before identification even begins. Budget six to eight weeks for a modest assemblage if you are doing it in house. Contract labs can move faster but they charge for it.
The field season itself is shorter than most people expect. You lose days to weather, equipment failure, and permitting delays. A two-week season might give you ten productive days if you are lucky. Plan accordingly. Underestimating the time needed for documentation is the most common scheduling error I see. People think they can dig and record in the same timeframe. They cannot. Recording takes at least as long as the excavation itself, usually longer.
Practical Considerations
Equipment costs add up. A good trowel costs twenty dollars. A reliable total station runs anywhere from three to ten thousand depending on specs. Sieves, screening frames, flotation columns, GPS units, and field computers are all line items. If you are working on a grant, include equipment in your budget from the start. Borrowing from another team is fine occasionally but not a sustainable strategy. Volunteer management is another practical concern. Trained volunteers can be extremely helpful for screening and surface collection. Untrained volunteers digging in sensitive areas is a risk you do not need. Invest time in training before anyone touches a trowel. A half-day orientation covering unit layout, trowel technique, and how to spot artifacts will prevent more problems than almost anything else you can do. Safety is straightforward but often ignored. Heat exhaustion, snake bites, ticks, and repetitive strain from years of bending over are the real hazards. Sun protection and hydration matter more than people realize in excavation contexts. I have seen experienced field directors push through heat illness because they did not want to lose a day of digging. It was a bad call. Nobody cares about your unit schedule if someone is in the ambulance.
When Methods Fail
Not every site yields to standard approaches. Dense urban settings with repeated construction activity can make stratigraphy nearly impossible to read. Coastal sites face erosion and salinity issues that degrade organic material rapidly. Tropical environments accelerate decomposition to the point that bone and wood may not survive at all. In those cases, you adapt your methods or accept that your research questions need to shift toward what can survive. There is no point forcing a method that the site conditions will not support. Another limitation is scale. Large landscape surveys using pedestrian transects and test pits are labor-intensive. A single researcher can cover maybe two hectares per day under good conditions. Anything larger requires a team and a longer timeframe. Drone-based photogrammetry helps with mapping but does not replace ground-truthing. You still need to walk the ground to verify anomalies and collect samples. The technology speeds things up but does not eliminate the physical work. Publishing standards continue to rise. Journals now expect data archiving, supplementary materials, and often raw spatial data. Plan for this from the beginning. Organizing your files properly during the project is dramatically easier than scrambling after you have left the field. I use a folder structure based on context number, date, and data type. It is simple and it scales. Spending an hour setting up that system at the start of a project pays for itself many times over.
The work is tedious and often frustrating. The ground does not cooperate. Equipment breaks. Weather ruins plans. But the alternative to doing it carefully is producing data that cannot be trusted, and that wastes everyone involved including the people who funded the project. Good archaeology is mostly good housekeeping with a little bit of theory attached to it.
